test suite: refactor: make initial_tenant optional

This patch makes the creation of the initial tenant & timeline optional. The motivation was https://github.com/neondatabase/neon/pull/3905#discussion_r1153627388 However, most existing tests require the env.initial_tenant. We don't want to change all of these here. So, the apporach taken here is to make NeonEnv a subclass of NeonEnvWithoutInitialTenant. It's not a proper subclass, but more of wrapper type. Hence the __new__ hackery. I guess this is a bad substitute for a Rust newtype + Deref impl. The initial tenant & timeline are created in NeonEnv, and the tenant is set as default (`--set-default`). We rely more on that default than before. Specifically, all neon_local invocation that previously used "--tenant-id", (tenant_id if tenant_id is not None else env.initial_tenant) now do *(["--tenant-id", tenant_id] if tenant_id is not None else []) The only real trouble with that was `Postgres` class's `pgdata_dir` member. I solved this by making `neon_local pg create` return the PGDATA dir path, instead of pre-computing it in Python. A net win, I think.
2026-07-11 16:10:36 +00:00 · 2023-03-31 01:00:59 +02:00
204 changed files with 3081 additions and 7587 deletions
--- a/.github/PULL_REQUEST_TEMPLATE/release-pr.md
+++ b/.github/PULL_REQUEST_TEMPLATE/release-pr.md
@@ -10,7 +10,6 @@
 <!-- List everything that should be done **before** release, any issues / setting changes / etc -->

 ### Checklist after release
- [ ] Make sure instructions from PRs included in this release and labeled `manual_release_instructions` are executed (either by you or by people who wrote them).
 - [ ] Based on the merged commits write release notes and open a PR into `website` repo ([example](https://github.com/neondatabase/website/pull/219/files))
 - [ ] Check [#dev-production-stream](https://neondb.slack.com/archives/C03F5SM1N02) Slack channel
 - [ ] Check [stuck projects page](https://console.neon.tech/admin/projects?sort=last_active&order=desc&stuck=true)
--- a/.github/actions/allure-report/action.yml
+++ b/.github/actions/allure-report/action.yml
@@ -15,42 +15,20 @@ outputs:
  report-url:
    description: 'Allure report URL'
    value: ${{ steps.generate-report.outputs.report-url }}
-  report-json-url:
-    description: 'Allure report JSON URL'
-    value: ${{ steps.generate-report.outputs.report-json-url }}

 runs:
  using: "composite"
-
  steps:
-    # We're using some of env variables quite offen, so let's set them once.
-    #
-    # It would be nice to have them set in common runs.env[0] section, but it doesn't work[1]
-    #
-    # - [0] https://docs.github.com/en/actions/creating-actions/metadata-syntax-for-github-actions#runsenv
-    # - [1] https://github.com/neondatabase/neon/pull/3907#discussion_r1154703456
-    #
-    - name: Set common environment variables
-      shell: bash -euxo pipefail {0}
-      run: |
-        echo "BUILD_TYPE=${BUILD_TYPE}"   >> $GITHUB_ENV
-        echo "BUCKET=${BUCKET}"           >> $GITHUB_ENV
-        echo "TEST_OUTPUT=${TEST_OUTPUT}" >> $GITHUB_ENV
-      env:
-        BUILD_TYPE: ${{ inputs.build_type }}
-        BUCKET: neon-github-public-dev
-        TEST_OUTPUT: /tmp/test_output
-
    - name: Validate input parameters
      shell: bash -euxo pipefail {0}
      run: |
        if [ "${{ inputs.action }}" != "store" ] && [ "${{ inputs.action }}" != "generate" ]; then
-          echo >&2 "Unknown inputs.action type '${{ inputs.action }}'; allowed 'generate' or 'store' only"
+          echo 2>&1 "Unknown inputs.action type '${{ inputs.action }}'; allowed 'generate' or 'store' only"
          exit 1
        fi

        if [ -z "${{ inputs.test_selection }}" ] && [ "${{ inputs.action }}" == "store" ]; then
-          echo >&2 "inputs.test_selection must be set for 'store' action"
+          echo 2>&1 "inputs.test_selection must be set for 'store' action"
          exit 2
        fi

@@ -98,14 +76,16 @@ runs:
          rm -f ${ALLURE_ZIP}
        fi
      env:
-        ALLURE_VERSION: 2.21.0
-        ALLURE_ZIP_MD5: c8db4dd8e2a7882583d569ed2c82879c
+        ALLURE_VERSION: 2.19.0
+        ALLURE_ZIP_MD5: ced21401a1a8b9dfb68cee9e4c210464

    - name: Upload Allure results
      if: ${{ inputs.action == 'store' }}
      env:
        REPORT_PREFIX: reports/${{ steps.calculate-vars.outputs.KEY }}/${{ inputs.build_type }}
        RAW_PREFIX: reports-raw/${{ steps.calculate-vars.outputs.KEY }}/${{ inputs.build_type }}
+        TEST_OUTPUT: /tmp/test_output
+        BUCKET: neon-github-public-dev
        TEST_SELECTION: ${{ steps.calculate-vars.outputs.TEST_SELECTION }}
      shell: bash -euxo pipefail {0}
      run: |
@@ -124,7 +104,7 @@ runs:
        EOF
        cat <<EOF > $TEST_OUTPUT/allure/results/environment.properties
          TEST_SELECTION=${{ inputs.test_selection }}
-          BUILD_TYPE=${BUILD_TYPE}
+          BUILD_TYPE=${{ inputs.build_type }}
        EOF

        ARCHIVE="${GITHUB_RUN_ID}-${TEST_SELECTION}-${GITHUB_RUN_ATTEMPT}-$(date +%s).tar.zst"
@@ -133,12 +113,13 @@ runs:
        tar -C ${TEST_OUTPUT}/allure/results -cf ${ARCHIVE} --zstd .
        aws s3 mv --only-show-errors ${ARCHIVE} "s3://${BUCKET}/${RAW_PREFIX}/${ARCHIVE}"

-    # Potentially we could have several running build for the same key (for example for the main branch), so we use improvised lock for this
+    # Potentially we could have several running build for the same key (for example for the main branch),  so we use improvised lock for this
    - name: Acquire Allure lock
      if: ${{ inputs.action == 'generate' }}
      shell: bash -euxo pipefail {0}
      env:
        LOCK_FILE: reports/${{ steps.calculate-vars.outputs.KEY }}/lock.txt
+        BUCKET: neon-github-public-dev
        TEST_SELECTION: ${{ steps.calculate-vars.outputs.TEST_SELECTION }}
      run: |
        LOCK_TIMEOUT=300 # seconds
@@ -168,6 +149,8 @@ runs:
      env:
        REPORT_PREFIX: reports/${{ steps.calculate-vars.outputs.KEY }}/${{ inputs.build_type }}
        RAW_PREFIX: reports-raw/${{ steps.calculate-vars.outputs.KEY }}/${{ inputs.build_type }}
+        TEST_OUTPUT: /tmp/test_output
+        BUCKET: neon-github-public-dev
      shell: bash -euxo pipefail {0}
      run: |
        # Get previously uploaded data for this run
@@ -203,24 +186,24 @@ runs:
        REPORT_URL=https://${BUCKET}.s3.amazonaws.com/${REPORT_PREFIX}/${GITHUB_RUN_ID}/index.html

        # Generate redirect
-        cat <<EOF > ${TEST_OUTPUT}/allure/index.html
+        cat <<EOF > ./index.html
          <!DOCTYPE html>

          <meta charset="utf-8">
          <title>Redirecting to ${REPORT_URL}</title>
          <meta http-equiv="refresh" content="0; URL=${REPORT_URL}">
        EOF
-        aws s3 cp --only-show-errors ${TEST_OUTPUT}/allure/index.html "s3://${BUCKET}/${REPORT_PREFIX}/latest/index.html"
+        aws s3 cp --only-show-errors ./index.html "s3://${BUCKET}/${REPORT_PREFIX}/latest/index.html"

        echo "[Allure Report](${REPORT_URL})" >> ${GITHUB_STEP_SUMMARY}
        echo "report-url=${REPORT_URL}" >> $GITHUB_OUTPUT
-        echo "report-json-url=${REPORT_URL%/index.html}/data/suites.json" >> $GITHUB_OUTPUT

    - name: Release Allure lock
      if: ${{ inputs.action == 'generate' && always() }}
      shell: bash -euxo pipefail {0}
      env:
        LOCK_FILE: reports/${{ steps.calculate-vars.outputs.KEY }}/lock.txt
+        BUCKET: neon-github-public-dev
        TEST_SELECTION: ${{ steps.calculate-vars.outputs.TEST_SELECTION }}
      run: |
        aws s3 cp --only-show-errors "s3://${BUCKET}/${LOCK_FILE}" ./lock.txt || exit 0
@@ -229,16 +212,11 @@ runs:
          aws s3 rm "s3://${BUCKET}/${LOCK_FILE}"
        fi

-    - name: Cleanup
-      if: always()
-      shell: bash -euxo pipefail {0}
-      run: |
-        rm -rf ${TEST_OUTPUT}/allure
-
    - uses: actions/github-script@v6
      if: ${{ inputs.action == 'generate' && always() }}
      env:
        REPORT_URL: ${{ steps.generate-report.outputs.report-url }}
+        BUILD_TYPE: ${{ inputs.build_type }}
        SHA: ${{ github.event.pull_request.head.sha || github.sha }}
      with:
        script: |
--- a/.github/actions/download/action.yml
+++ b/.github/actions/download/action.yml
@@ -37,7 +37,7 @@ runs:
            echo 'SKIPPED=true' >> $GITHUB_OUTPUT
            exit 0
          else
-            echo >&2 "Neither s3://${BUCKET}/${PREFIX}/${FILENAME} nor its version from previous attempts exist"
+            echo 2>&1 "Neither s3://${BUCKET}/${PREFIX}/${FILENAME} nor its version from previous attempts exist"
            exit 1
          fi
        fi
--- a/.github/actions/neon-branch-create/action.yml
+++ b/.github/actions/neon-branch-create/action.yml
@@ -58,7 +58,7 @@ runs:
        done

        if [ -z "${branch_id}" ] || [ "${branch_id}" == "null" ]; then
-          echo >&2 "Failed to create branch after 10 attempts, the latest response was: ${branch}"
+          echo 2>&1 "Failed to create branch after 10 attempts, the latest response was: ${branch}"
          exit 1
        fi

@@ -122,7 +122,7 @@ runs:
        done

        if [ -z "${password}" ] || [ "${password}" == "null" ]; then
-          echo >&2 "Failed to reset password after 10 attempts, the latest response was: ${reset_password}"
+          echo 2>&1 "Failed to reset password after 10 attempts, the latest response was: ${reset_password}"
          exit 1
        fi

--- a/.github/actions/neon-branch-delete/action.yml
+++ b/.github/actions/neon-branch-delete/action.yml
@@ -48,7 +48,7 @@ runs:
        done

        if [ -z "${branch_id}" ] || [ "${branch_id}" == "null" ]; then
-          echo >&2 "Failed to delete branch after 10 attempts, the latest response was: ${deleted_branch}"
+          echo 2>&1 "Failed to delete branch after 10 attempts, the latest response was: ${deleted_branch}"
          exit 1
        fi
      env:
--- a/.github/actions/neon-project-create/action.yml
+++ b/.github/actions/neon-project-create/action.yml
@@ -14,12 +14,6 @@ inputs:
  api_host:
    desctiption: 'Neon API host'
    default: console.stage.neon.tech
-  provisioner:
-    desctiption: 'k8s-pod or k8s-neonvm'
-    default: 'k8s-pod'
-  compute_units:
-    desctiption: '[Min, Max] compute units; Min and Max are used for k8s-neonvm with autoscaling, for k8s-pod values Min and Max should be equal'
-    default: '[1, 1]'

 outputs:
  dsn:
@@ -37,10 +31,6 @@ runs:
      # A shell without `set -x` to not to expose password/dsn in logs
      shell: bash -euo pipefail {0}
      run: |
-        if [ "${PROVISIONER}" == "k8s-pod" ] && [ "${MIN_CU}" != "${MAX_CU}" ]; then
-          echo >&2 "For k8s-pod provisioner MIN_CU should be equal to MAX_CU"
-        fi
-
        project=$(curl \
          "https://${API_HOST}/api/v2/projects" \
          --fail \
@@ -52,9 +42,6 @@ runs:
              \"name\": \"Created by actions/neon-project-create; GITHUB_RUN_ID=${GITHUB_RUN_ID}\",
              \"pg_version\": ${POSTGRES_VERSION},
              \"region_id\": \"${REGION_ID}\",
-              \"provisioner\": \"${PROVISIONER}\",
-              \"autoscaling_limit_min_cu\": ${MIN_CU},
-              \"autoscaling_limit_max_cu\": ${MAX_CU},
              \"settings\": { }
            }
          }")
@@ -75,6 +62,3 @@ runs:
        API_KEY: ${{ inputs.api_key }}
        REGION_ID: ${{ inputs.region_id }}
        POSTGRES_VERSION: ${{ inputs.postgres_version }}
-        PROVISIONER: ${{ inputs.provisioner }}
-        MIN_CU: ${{ fromJSON(inputs.compute_units)[0] }}
-        MAX_CU: ${{ fromJSON(inputs.compute_units)[1] }}
--- a/.github/actions/run-python-test-set/action.yml
+++ b/.github/actions/run-python-test-set/action.yml
@@ -44,10 +44,6 @@ inputs:
    description: 'Secret access key'
    required: false
    default: ''
-  rerun_flaky:
-    description: 'Whether to rerun flaky tests'
-    required: false
-    default: 'false'

 runs:
  using: "composite"
@@ -105,7 +101,6 @@ runs:
        COMPATIBILITY_SNAPSHOT_DIR: /tmp/compatibility_snapshot_pg14
        ALLOW_BACKWARD_COMPATIBILITY_BREAKAGE: contains(github.event.pull_request.labels.*.name, 'backward compatibility breakage')
        ALLOW_FORWARD_COMPATIBILITY_BREAKAGE: contains(github.event.pull_request.labels.*.name, 'forward compatibility breakage')
-        RERUN_FLAKY: ${{ inputs.rerun_flaky }}
      shell: bash -euxo pipefail {0}
      run: |
        # PLATFORM will be embedded in the perf test report
@@ -148,13 +143,6 @@ runs:
          EXTRA_PARAMS="--out-dir $PERF_REPORT_DIR $EXTRA_PARAMS"
        fi

-        if [ "${RERUN_FLAKY}" == "true" ]; then
-          mkdir -p $TEST_OUTPUT
-          poetry run ./scripts/flaky_tests.py "${TEST_RESULT_CONNSTR}" --days 10 --output "$TEST_OUTPUT/flaky.json"
-
-          EXTRA_PARAMS="--flaky-tests-json $TEST_OUTPUT/flaky.json $EXTRA_PARAMS"
-        fi
-
        if [[ "${{ inputs.build_type }}" == "debug" ]]; then
          cov_prefix=(scripts/coverage "--profraw-prefix=$GITHUB_JOB" --dir=/tmp/coverage run)
        elif [[ "${{ inputs.build_type }}" == "release" ]]; then
@@ -202,7 +190,7 @@ runs:
        prefix: latest

    - name: Create Allure report
-      if: ${{ !cancelled() }}
+      if: success() || failure()
      uses: ./.github/actions/allure-report
      with:
        action: store
--- a/.github/actions/upload/action.yml
+++ b/.github/actions/upload/action.yml
@@ -23,7 +23,7 @@ runs:
        mkdir -p $(dirname $ARCHIVE)

        if [ -f ${ARCHIVE} ]; then
-          echo >&2 "File ${ARCHIVE} already exist. Something went wrong before"
+          echo 2>&1 "File ${ARCHIVE} already exist. Something went wrong before"
          exit 1
        fi

@@ -33,10 +33,10 @@ runs:
        elif [ -f ${SOURCE} ]; then
          time tar -cf ${ARCHIVE} --zstd ${SOURCE}
        elif ! ls ${SOURCE} > /dev/null 2>&1; then
-          echo >&2 "${SOURCE} does not exist"
+          echo 2>&1 "${SOURCE} does not exist"
          exit 2
        else
-          echo >&2 "${SOURCE} is neither a directory nor a file, do not know how to handle it"
+          echo 2>&1 "${SOURCE} is neither a directory nor a file, do not know how to handle it"
          exit 3
        fi

--- a/.github/ansible/prod.ap-southeast-1.hosts.yaml
+++ b/.github/ansible/prod.ap-southeast-1.hosts.yaml
@@ -8,16 +8,6 @@ storage:
      pg_distrib_dir: /usr/local
      metric_collection_endpoint: http://neon-internal-api.aws.neon.tech/billing/api/v1/usage_events
      metric_collection_interval: 10min
-      disk_usage_based_eviction:
-        max_usage_pct: 85 # TODO: decrease to 80 after all pageservers are below 80
-        min_avail_bytes: 0
-        period: "10s"
-      tenant_config:
-        eviction_policy:
-          kind: "LayerAccessThreshold"
-          period: "10m"
-          threshold: &default_eviction_threshold "24h"
-      evictions_low_residence_duration_metric_threshold: *default_eviction_threshold
      remote_storage:
        bucket_name: "{{ bucket_name }}"
        bucket_region: "{{ bucket_region }}"
--- a/.github/ansible/prod.eu-central-1.hosts.yaml
+++ b/.github/ansible/prod.eu-central-1.hosts.yaml
@@ -8,16 +8,6 @@ storage:
      pg_distrib_dir: /usr/local
      metric_collection_endpoint: http://neon-internal-api.aws.neon.tech/billing/api/v1/usage_events
      metric_collection_interval: 10min
-      disk_usage_based_eviction:
-        max_usage_pct: 85 # TODO: decrease to 80 after all pageservers are below 80
-        min_avail_bytes: 0
-        period: "10s"
-      tenant_config:
-        eviction_policy:
-          kind: "LayerAccessThreshold"
-          period: "10m"
-          threshold: &default_eviction_threshold "24h"
-      evictions_low_residence_duration_metric_threshold: *default_eviction_threshold
      remote_storage:
        bucket_name: "{{ bucket_name }}"
        bucket_region: "{{ bucket_region }}"
--- a/.github/ansible/prod.us-east-2.hosts.yaml
+++ b/.github/ansible/prod.us-east-2.hosts.yaml
@@ -8,16 +8,6 @@ storage:
      pg_distrib_dir: /usr/local
      metric_collection_endpoint: http://neon-internal-api.aws.neon.tech/billing/api/v1/usage_events
      metric_collection_interval: 10min
-      disk_usage_based_eviction:
-        max_usage_pct: 85 # TODO: decrease to 80 after all pageservers are below 80
-        min_avail_bytes: 0
-        period: "10s"
-      tenant_config:
-        eviction_policy:
-          kind: "LayerAccessThreshold"
-          period: "10m"
-          threshold: &default_eviction_threshold "24h"
-      evictions_low_residence_duration_metric_threshold: *default_eviction_threshold
      remote_storage:
        bucket_name: "{{ bucket_name }}"
        bucket_region: "{{ bucket_region }}"
--- a/.github/ansible/prod.us-west-2.hosts.yaml
+++ b/.github/ansible/prod.us-west-2.hosts.yaml
@@ -8,16 +8,6 @@ storage:
      pg_distrib_dir: /usr/local
      metric_collection_endpoint: http://neon-internal-api.aws.neon.tech/billing/api/v1/usage_events
      metric_collection_interval: 10min
-      disk_usage_based_eviction:
-        max_usage_pct: 85 # TODO: decrease to 80 after all pageservers are below 80
-        min_avail_bytes: 0
-        period: "10s"
-      tenant_config:
-        eviction_policy:
-          kind: "LayerAccessThreshold"
-          period: "10m"
-          threshold: &default_eviction_threshold "24h"
-      evictions_low_residence_duration_metric_threshold: *default_eviction_threshold
      remote_storage:
        bucket_name: "{{ bucket_name }}"
        bucket_region: "{{ bucket_region }}"
--- a/.github/ansible/scripts/init_pageserver.sh
+++ b/.github/ansible/scripts/init_pageserver.sh
@@ -3,8 +3,6 @@
 # fetch params from meta-data service
 INSTANCE_ID=$(curl -s http://169.254.169.254/latest/meta-data/instance-id)
 AZ_ID=$(curl -s http://169.254.169.254/latest/meta-data/placement/availability-zone)
-INSTANCE_TYPE=$(curl -s http://169.254.169.254/latest/meta-data/instance-type)
-DISK_SIZE=$(df -B1 /storage | tail -1 | awk '{print $2}')

 # store fqdn hostname in var
 HOST=$(hostname -f)
@@ -20,9 +18,7 @@ cat <<EOF | tee /tmp/payload
  "http_host": "${HOST}",
  "http_port": 9898,
  "active": false,
-  "availability_zone_id": "${AZ_ID}",
-  "disk_size": ${DISK_SIZE},
-  "instance_type": "${INSTANCE_TYPE}"
+  "availability_zone_id": "${AZ_ID}"
 }
 EOF

--- a/.github/ansible/staging.eu-west-1.hosts.yaml
+++ b/.github/ansible/staging.eu-west-1.hosts.yaml
@@ -8,16 +8,11 @@ storage:
      pg_distrib_dir: /usr/local
      metric_collection_endpoint: http://neon-internal-api.aws.neon.build/billing/api/v1/usage_events
      metric_collection_interval: 10min
-      disk_usage_based_eviction:
-        max_usage_pct: 80
-        min_avail_bytes: 0
-        period: "10s"
      tenant_config:
        eviction_policy:
          kind: "LayerAccessThreshold"
          period: "20m"
-          threshold: &default_eviction_threshold "20m"
-      evictions_low_residence_duration_metric_threshold: *default_eviction_threshold
+          threshold: "20m"
      remote_storage:
        bucket_name: "{{ bucket_name }}"
        bucket_region: "{{ bucket_region }}"
--- a/.github/ansible/staging.us-east-2.hosts.yaml
+++ b/.github/ansible/staging.us-east-2.hosts.yaml
@@ -8,16 +8,11 @@ storage:
      pg_distrib_dir: /usr/local
      metric_collection_endpoint: http://neon-internal-api.aws.neon.build/billing/api/v1/usage_events
      metric_collection_interval: 10min
-      disk_usage_based_eviction:
-        max_usage_pct: 80
-        min_avail_bytes: 0
-        period: "10s"
      tenant_config:
        eviction_policy:
          kind: "LayerAccessThreshold"
          period: "20m"
-          threshold: &default_eviction_threshold "20m"
-      evictions_low_residence_duration_metric_threshold: *default_eviction_threshold
+          threshold: "20m"
      remote_storage:
        bucket_name: "{{ bucket_name }}"
        bucket_region: "{{ bucket_region }}"
--- a/.github/helm-values/dev-us-east-2-beta.neon-proxy-scram.yaml
+++ b/.github/helm-values/dev-us-east-2-beta.neon-proxy-scram.yaml
@@ -23,7 +23,6 @@ settings:
  authBackend: "console"
  authEndpoint: "http://neon-internal-api.aws.neon.build/management/api/v2"
  domain: "*.us-east-2.aws.neon.build"
-  extraDomains: ["*.us-east-2.postgres.zenith.tech", "*.us-east-2.retooldb-staging.com"]
  sentryEnvironment: "staging"
  wssPort: 8443
  metricCollectionEndpoint: "http://neon-internal-api.aws.neon.build/billing/api/v1/usage_events"
--- a/.github/helm-values/prod-ap-southeast-1-epsilon.neon-proxy-scram.yaml
+++ b/.github/helm-values/prod-ap-southeast-1-epsilon.neon-proxy-scram.yaml
@@ -24,7 +24,6 @@ settings:
  authBackend: "console"
  authEndpoint: "http://neon-internal-api.aws.neon.tech/management/api/v2"
  domain: "*.ap-southeast-1.aws.neon.tech"
-  extraDomains: ["*.ap-southeast-1.retooldb.com", "*.ap-southeast-1.postgres.vercel-storage.com"]
  sentryEnvironment: "production"
  wssPort: 8443
  metricCollectionEndpoint: "http://neon-internal-api.aws.neon.tech/billing/api/v1/usage_events"
--- a/.github/helm-values/prod-eu-central-1-gamma.neon-proxy-scram.yaml
+++ b/.github/helm-values/prod-eu-central-1-gamma.neon-proxy-scram.yaml
@@ -24,7 +24,6 @@ settings:
  authBackend: "console"
  authEndpoint: "http://neon-internal-api.aws.neon.tech/management/api/v2"
  domain: "*.eu-central-1.aws.neon.tech"
-  extraDomains: ["*.eu-central-1.retooldb.com", "*.eu-central-1.postgres.vercel-storage.com"]
  sentryEnvironment: "production"
  wssPort: 8443
  metricCollectionEndpoint: "http://neon-internal-api.aws.neon.tech/billing/api/v1/usage_events"
--- a/.github/helm-values/prod-us-east-2-delta.neon-proxy-scram.yaml
+++ b/.github/helm-values/prod-us-east-2-delta.neon-proxy-scram.yaml
@@ -24,7 +24,6 @@ settings:
  authBackend: "console"
  authEndpoint: "http://neon-internal-api.aws.neon.tech/management/api/v2"
  domain: "*.us-east-2.aws.neon.tech"
-  extraDomains: ["*.us-east-2.retooldb.com", "*.us-east-2.postgres.vercel-storage.com"]
  sentryEnvironment: "production"
  wssPort: 8443
  metricCollectionEndpoint: "http://neon-internal-api.aws.neon.tech/billing/api/v1/usage_events"
--- a/.github/helm-values/prod-us-west-2-eta.neon-proxy-scram.yaml
+++ b/.github/helm-values/prod-us-west-2-eta.neon-proxy-scram.yaml
@@ -24,7 +24,6 @@ settings:
  authBackend: "console"
  authEndpoint: "http://neon-internal-api.aws.neon.tech/management/api/v2"
  domain: "*.us-west-2.aws.neon.tech"
-  extraDomains: ["*.us-west-2.retooldb.com", "*.us-west-2.postgres.vercel-storage.com"]
  sentryEnvironment: "production"
  wssPort: 8443
  metricCollectionEndpoint: "http://neon-internal-api.aws.neon.tech/billing/api/v1/usage_events"
--- a/.github/pull_request_template.md
+++ b/.github/pull_request_template.md
@@ -3,12 +3,8 @@
 ## Issue ticket number and link

 ## Checklist before requesting a review
-
 - [ ] I have performed a self-review of my code.
 - [ ] If it is a core feature, I have added thorough tests.
 - [ ] Do we need to implement analytics? if so did you add the relevant metrics to the dashboard?
 - [ ] If this PR requires public announcement, mark it with /release-notes label and add several sentences in this section.

-## Checklist before merging
-
- [ ] Do not forget to reformat commit message to not include the above checklist
--- a/.github/workflows/benchmarking.yml
+++ b/.github/workflows/benchmarking.yml
@@ -30,7 +30,7 @@ defaults:

 concurrency:
  # Allow only one workflow per any non-`main` branch.
-  group: ${{ github.workflow }}-${{ github.ref_name }}-${{ github.ref_name == 'main' && github.sha || 'anysha' }}
+  group: ${{ github.workflow }}-${{ github.ref }}-${{ github.ref == 'refs/heads/main' && github.sha || 'anysha' }}
  cancel-in-progress: true

 jobs:
@@ -42,7 +42,7 @@ jobs:
      DEFAULT_PG_VERSION: 14
      TEST_OUTPUT: /tmp/test_output
      BUILD_TYPE: remote
-      SAVE_PERF_REPORT: ${{ github.event.inputs.save_perf_report || ( github.ref_name == 'main' ) }}
+      SAVE_PERF_REPORT: ${{ github.event.inputs.save_perf_report || ( github.ref == 'refs/heads/main' ) }}
      PLATFORM: "neon-staging"

    runs-on: [ self-hosted, us-east-2, x64 ]
@@ -92,7 +92,7 @@ jobs:
        api_key: ${{ secrets.NEON_STAGING_API_KEY }}

    - name: Create Allure report
-      if: ${{ !cancelled() }}
+      if: success() || failure()
      uses: ./.github/actions/allure-report
      with:
        action: generate
@@ -107,65 +107,25 @@ jobs:
      env:
        SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}

-  generate-matrices:
-    # Create matrices for the benchmarking jobs, so we run benchmarks on rds only once a week (on Saturday)
-    #
-    # Available platforms:
-    # - neon-captest-new: Freshly created project (1 CU)
-    # - neon-captest-freetier: Use freetier-sized compute (0.25 CU)
-    # - neon-captest-reuse: Reusing existing project
-    # - rds-aurora: Aurora Postgres Serverless v2 with autoscaling from 0.5 to 2 ACUs
-    # - rds-postgres: RDS Postgres db.m5.large instance (2 vCPU, 8 GiB) with gp3 EBS storage
-    runs-on: ubuntu-latest
-    outputs:
-      pgbench-compare-matrix: ${{ steps.pgbench-compare-matrix.outputs.matrix }}
-      olap-compare-matrix: ${{ steps.olap-compare-matrix.outputs.matrix }}
-
-    steps:
-    - name: Generate matrix for pgbench benchmark
-      id: pgbench-compare-matrix
-      run: |
-        matrix='{
-          "platform": [
-            "neon-captest-new",
-            "neon-captest-reuse"
-          ],
-          "db_size": [ "10gb" ],
-          "include": [
-            { "platform": "neon-captest-freetier", "db_size": "3gb"  },
-            { "platform": "neon-captest-new",      "db_size": "50gb" }
-          ]
-        }'
-
-        if [ "$(date +%A)" = "Saturday" ]; then
-          matrix=$(echo $matrix | jq '.include += [{ "platform": "rds-postgres", "db_size": "10gb"},
-                                                   { "platform": "rds-aurora",   "db_size": "50gb"}]')
-        fi
-
-        echo "matrix=$(echo $matrix | jq --compact-output '.')" >> $GITHUB_OUTPUT
-
-    - name: Generate matrix for OLAP benchmarks
-      id: olap-compare-matrix
-      run: |
-        matrix='{
-          "platform": [
-            "neon-captest-reuse"
-          ]
-        }'
-
-        if [ "$(date +%A)" = "Saturday" ]; then
-          matrix=$(echo $matrix | jq '.include += [{ "platform": "rds-postgres" },
-                                                   { "platform": "rds-aurora"   }]')
-        fi
-
-        echo "matrix=$(echo $matrix | jq --compact-output '.')" >> $GITHUB_OUTPUT
-
  pgbench-compare:
-    needs: [ generate-matrices ]
-
    strategy:
      fail-fast: false
-      matrix: ${{fromJson(needs.generate-matrices.outputs.pgbench-compare-matrix)}}
+      matrix:
+        # neon-captest-new: Run pgbench in a freshly created project
+        # neon-captest-reuse: Same, but reusing existing project
+        # neon-captest-prefetch: Same, with prefetching enabled (new project)
+        # rds-aurora: Aurora Postgres Serverless v2 with autoscaling from 0.5 to 2 ACUs
+        # rds-postgres: RDS Postgres db.m5.large instance (2 vCPU, 8 GiB) with gp3 EBS storage
+        platform: [ neon-captest-reuse, neon-captest-prefetch, rds-postgres ]
+        db_size: [ 10gb ]
+        runner: [ us-east-2 ]
+        include:
+          - platform: neon-captest-prefetch
+            db_size: 50gb
+            runner: us-east-2
+          - platform: rds-aurora
+            db_size: 50gb
+            runner: us-east-2

    env:
      TEST_PG_BENCH_DURATIONS_MATRIX: "60m"
@@ -174,10 +134,10 @@ jobs:
      DEFAULT_PG_VERSION: 14
      TEST_OUTPUT: /tmp/test_output
      BUILD_TYPE: remote
-      SAVE_PERF_REPORT: ${{ github.event.inputs.save_perf_report || ( github.ref_name == 'main' ) }}
+      SAVE_PERF_REPORT: ${{ github.event.inputs.save_perf_report || ( github.ref == 'refs/heads/main' ) }}
      PLATFORM: ${{ matrix.platform }}

-    runs-on: [ self-hosted, us-east-2, x64 ]
+    runs-on: [ self-hosted, "${{ matrix.runner }}", x64 ]
    container:
      image: 369495373322.dkr.ecr.eu-central-1.amazonaws.com/rust:pinned
      options: --init
@@ -200,14 +160,13 @@ jobs:
        echo "${POSTGRES_DISTRIB_DIR}/v${DEFAULT_PG_VERSION}/bin" >> $GITHUB_PATH

    - name: Create Neon Project
-      if: contains(fromJson('["neon-captest-new", "neon-captest-freetier"]'), matrix.platform)
+      if: contains(fromJson('["neon-captest-new", "neon-captest-prefetch"]'), matrix.platform)
      id: create-neon-project
      uses: ./.github/actions/neon-project-create
      with:
        region_id: ${{ github.event.inputs.region_id || 'aws-us-east-2' }}
        postgres_version: ${{ env.DEFAULT_PG_VERSION }}
        api_key: ${{ secrets.NEON_STAGING_API_KEY }}
-        compute_units: ${{ (matrix.platform == 'neon-captest-freetier' && '[0.25, 0.25]') || '[1, 1]' }}

    - name: Set up Connection String
      id: set-up-connstr
@@ -216,7 +175,7 @@ jobs:
          neon-captest-reuse)
            CONNSTR=${{ secrets.BENCHMARK_CAPTEST_CONNSTR }}
            ;;
-          neon-captest-new | neon-captest-freetier)
+          neon-captest-new | neon-captest-prefetch)
            CONNSTR=${{ steps.create-neon-project.outputs.dsn }}
            ;;
          rds-aurora)
@@ -226,7 +185,7 @@ jobs:
            CONNSTR=${{ secrets.BENCHMARK_RDS_POSTGRES_CONNSTR }}
            ;;
          *)
-            echo >&2 "Unknown PLATFORM=${PLATFORM}. Allowed only 'neon-captest-reuse', 'neon-captest-new', 'neon-captest-freetier', 'rds-aurora', or 'rds-postgres'"
+            echo 2>&1 "Unknown PLATFORM=${PLATFORM}. Allowed only 'neon-captest-reuse', 'neon-captest-new', 'neon-captest-prefetch', 'rds-aurora', or 'rds-postgres'"
            exit 1
            ;;
        esac
@@ -235,6 +194,17 @@ jobs:

        psql ${CONNSTR} -c "SELECT version();"

+    - name: Set database options
+      if: matrix.platform == 'neon-captest-prefetch'
+      run: |
+        DB_NAME=$(psql ${BENCHMARK_CONNSTR} --no-align --quiet -t -c "SELECT current_database()")
+
+        psql ${BENCHMARK_CONNSTR} -c "ALTER DATABASE ${DB_NAME} SET enable_seqscan_prefetch=on"
+        psql ${BENCHMARK_CONNSTR} -c "ALTER DATABASE ${DB_NAME} SET effective_io_concurrency=32"
+        psql ${BENCHMARK_CONNSTR} -c "ALTER DATABASE ${DB_NAME} SET maintenance_io_concurrency=32"
+      env:
+        BENCHMARK_CONNSTR: ${{ steps.set-up-connstr.outputs.connstr }}
+
    - name: Benchmark init
      uses: ./.github/actions/run-python-test-set
      with:
@@ -282,7 +252,7 @@ jobs:
        api_key: ${{ secrets.NEON_STAGING_API_KEY }}

    - name: Create Allure report
-      if: ${{ !cancelled() }}
+      if: success() || failure()
      uses: ./.github/actions/allure-report
      with:
        action: generate
@@ -305,19 +275,23 @@ jobs:
    #
    # *_CLICKBENCH_CONNSTR: Genuine ClickBench DB with ~100M rows
    # *_CLICKBENCH_10M_CONNSTR: DB with the first 10M rows of ClickBench DB
-    if: ${{ !cancelled() }}
-    needs: [ generate-matrices, pgbench-compare ]
+    if: success() || failure()
+    needs: [ pgbench-compare ]

    strategy:
      fail-fast: false
-      matrix: ${{ fromJson(needs.generate-matrices.outputs.olap-compare-matrix) }}
+      matrix:
+        # neon-captest-prefetch: We have pre-created projects with prefetch enabled
+        # rds-aurora: Aurora Postgres Serverless v2 with autoscaling from 0.5 to 2 ACUs
+        # rds-postgres: RDS Postgres db.m5.large instance (2 vCPU, 8 GiB) with gp3 EBS storage
+        platform: [ neon-captest-prefetch, rds-postgres, rds-aurora ]

    env:
      POSTGRES_DISTRIB_DIR: /tmp/neon/pg_install
      DEFAULT_PG_VERSION: 14
      TEST_OUTPUT: /tmp/test_output
      BUILD_TYPE: remote
-      SAVE_PERF_REPORT: ${{ github.event.inputs.save_perf_report || ( github.ref_name == 'main' ) }}
+      SAVE_PERF_REPORT: ${{ github.event.inputs.save_perf_report || ( github.ref == 'refs/heads/main' ) }}
      PLATFORM: ${{ matrix.platform }}

    runs-on: [ self-hosted, us-east-2, x64 ]
@@ -346,7 +320,7 @@ jobs:
      id: set-up-connstr
      run: |
        case "${PLATFORM}" in
-          neon-captest-reuse)
+          neon-captest-prefetch)
            CONNSTR=${{ secrets.BENCHMARK_CAPTEST_CLICKBENCH_10M_CONNSTR }}
            ;;
          rds-aurora)
@@ -356,7 +330,7 @@ jobs:
            CONNSTR=${{ secrets.BENCHMARK_RDS_POSTGRES_CLICKBENCH_10M_CONNSTR }}
            ;;
          *)
-            echo >&2 "Unknown PLATFORM=${PLATFORM}. Allowed only 'neon-captest-reuse', 'rds-aurora', or 'rds-postgres'"
+            echo 2>&1 "Unknown PLATFORM=${PLATFORM}. Allowed only 'neon-captest-prefetch', 'rds-aurora', or 'rds-postgres'"
            exit 1
            ;;
        esac
@@ -365,6 +339,17 @@ jobs:

        psql ${CONNSTR} -c "SELECT version();"

+    - name: Set database options
+      if: matrix.platform == 'neon-captest-prefetch'
+      run: |
+        DB_NAME=$(psql ${BENCHMARK_CONNSTR} --no-align --quiet -t -c "SELECT current_database()")
+
+        psql ${BENCHMARK_CONNSTR} -c "ALTER DATABASE ${DB_NAME} SET enable_seqscan_prefetch=on"
+        psql ${BENCHMARK_CONNSTR} -c "ALTER DATABASE ${DB_NAME} SET effective_io_concurrency=32"
+        psql ${BENCHMARK_CONNSTR} -c "ALTER DATABASE ${DB_NAME} SET maintenance_io_concurrency=32"
+      env:
+        BENCHMARK_CONNSTR: ${{ steps.set-up-connstr.outputs.connstr }}
+
    - name: ClickBench benchmark
      uses: ./.github/actions/run-python-test-set
      with:
@@ -379,7 +364,7 @@ jobs:
        BENCHMARK_CONNSTR: ${{ steps.set-up-connstr.outputs.connstr }}

    - name: Create Allure report
-      if: ${{ !cancelled() }}
+      if: success() || failure()
      uses: ./.github/actions/allure-report
      with:
        action: generate
@@ -401,19 +386,23 @@ jobs:
    # We might change it after https://github.com/neondatabase/neon/issues/2900.
    #
    # *_TPCH_S10_CONNSTR: DB generated with scale factor 10 (~10 GB)
-    if: ${{ !cancelled() }}
-    needs: [ generate-matrices, clickbench-compare ]
+    if: success() || failure()
+    needs: [ clickbench-compare ]

    strategy:
      fail-fast: false
-      matrix: ${{ fromJson(needs.generate-matrices.outputs.olap-compare-matrix) }}
+      matrix:
+        # neon-captest-prefetch: We have pre-created projects with prefetch enabled
+        # rds-aurora: Aurora Postgres Serverless v2 with autoscaling from 0.5 to 2 ACUs
+        # rds-postgres: RDS Postgres db.m5.large instance (2 vCPU, 8 GiB) with gp3 EBS storage
+        platform: [ neon-captest-prefetch, rds-postgres, rds-aurora ]

    env:
      POSTGRES_DISTRIB_DIR: /tmp/neon/pg_install
      DEFAULT_PG_VERSION: 14
      TEST_OUTPUT: /tmp/test_output
      BUILD_TYPE: remote
-      SAVE_PERF_REPORT: ${{ github.event.inputs.save_perf_report || ( github.ref_name == 'main' ) }}
+      SAVE_PERF_REPORT: ${{ github.event.inputs.save_perf_report || ( github.ref == 'refs/heads/main' ) }}
      PLATFORM: ${{ matrix.platform }}

    runs-on: [ self-hosted, us-east-2, x64 ]
@@ -442,7 +431,7 @@ jobs:
      id: set-up-connstr
      run: |
        case "${PLATFORM}" in
-          neon-captest-reuse)
+          neon-captest-prefetch)
            CONNSTR=${{ secrets.BENCHMARK_CAPTEST_TPCH_S10_CONNSTR }}
            ;;
          rds-aurora)
@@ -452,7 +441,7 @@ jobs:
            CONNSTR=${{ secrets.BENCHMARK_RDS_POSTGRES_TPCH_S10_CONNSTR }}
            ;;
          *)
-            echo >&2 "Unknown PLATFORM=${PLATFORM}. Allowed only 'neon-captest-reuse', 'rds-aurora', or 'rds-postgres'"
+            echo 2>&1 "Unknown PLATFORM=${PLATFORM}. Allowed only 'neon-captest-prefetch', 'rds-aurora', or 'rds-postgres'"
            exit 1
            ;;
        esac
@@ -461,6 +450,17 @@ jobs:

        psql ${CONNSTR} -c "SELECT version();"

+    - name: Set database options
+      if: matrix.platform == 'neon-captest-prefetch'
+      run: |
+        DB_NAME=$(psql ${BENCHMARK_CONNSTR} --no-align --quiet -t -c "SELECT current_database()")
+
+        psql ${BENCHMARK_CONNSTR} -c "ALTER DATABASE ${DB_NAME} SET enable_seqscan_prefetch=on"
+        psql ${BENCHMARK_CONNSTR} -c "ALTER DATABASE ${DB_NAME} SET effective_io_concurrency=32"
+        psql ${BENCHMARK_CONNSTR} -c "ALTER DATABASE ${DB_NAME} SET maintenance_io_concurrency=32"
+      env:
+        BENCHMARK_CONNSTR: ${{ steps.set-up-connstr.outputs.connstr }}
+
    - name: Run TPC-H benchmark
      uses: ./.github/actions/run-python-test-set
      with:
@@ -475,7 +475,7 @@ jobs:
        BENCHMARK_CONNSTR: ${{ steps.set-up-connstr.outputs.connstr }}

    - name: Create Allure report
-      if: ${{ !cancelled() }}
+      if: success() || failure()
      uses: ./.github/actions/allure-report
      with:
        action: generate
@@ -491,19 +491,23 @@ jobs:
        SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}

  user-examples-compare:
-    if: ${{ !cancelled() }}
-    needs: [ generate-matrices, tpch-compare ]
+    if: success() || failure()
+    needs: [ tpch-compare ]

    strategy:
      fail-fast: false
-      matrix: ${{ fromJson(needs.generate-matrices.outputs.olap-compare-matrix) }}
+      matrix:
+        # neon-captest-prefetch: We have pre-created projects with prefetch enabled
+        # rds-aurora: Aurora Postgres Serverless v2 with autoscaling from 0.5 to 2 ACUs
+        # rds-postgres: RDS Postgres db.m5.large instance (2 vCPU, 8 GiB) with gp3 EBS storage
+        platform: [ neon-captest-prefetch, rds-postgres, rds-aurora ]

    env:
      POSTGRES_DISTRIB_DIR: /tmp/neon/pg_install
      DEFAULT_PG_VERSION: 14
      TEST_OUTPUT: /tmp/test_output
      BUILD_TYPE: remote
-      SAVE_PERF_REPORT: ${{ github.event.inputs.save_perf_report || ( github.ref_name == 'main' ) }}
+      SAVE_PERF_REPORT: ${{ github.event.inputs.save_perf_report || ( github.ref == 'refs/heads/main' ) }}
      PLATFORM: ${{ matrix.platform }}

    runs-on: [ self-hosted, us-east-2, x64 ]
@@ -532,7 +536,7 @@ jobs:
      id: set-up-connstr
      run: |
        case "${PLATFORM}" in
-          neon-captest-reuse)
+          neon-captest-prefetch)
            CONNSTR=${{ secrets.BENCHMARK_USER_EXAMPLE_CAPTEST_CONNSTR }}
            ;;
          rds-aurora)
@@ -542,7 +546,7 @@ jobs:
            CONNSTR=${{ secrets.BENCHMARK_USER_EXAMPLE_RDS_POSTGRES_CONNSTR }}
            ;;
          *)
-            echo >&2 "Unknown PLATFORM=${PLATFORM}. Allowed only 'neon-captest-reuse', 'rds-aurora', or 'rds-postgres'"
+            echo 2>&1 "Unknown PLATFORM=${PLATFORM}. Allowed only 'neon-captest-prefetch', 'rds-aurora', or 'rds-postgres'"
            exit 1
            ;;
        esac
@@ -551,6 +555,17 @@ jobs:

        psql ${CONNSTR} -c "SELECT version();"

+    - name: Set database options
+      if: matrix.platform == 'neon-captest-prefetch'
+      run: |
+        DB_NAME=$(psql ${BENCHMARK_CONNSTR} --no-align --quiet -t -c "SELECT current_database()")
+
+        psql ${BENCHMARK_CONNSTR} -c "ALTER DATABASE ${DB_NAME} SET enable_seqscan_prefetch=on"
+        psql ${BENCHMARK_CONNSTR} -c "ALTER DATABASE ${DB_NAME} SET effective_io_concurrency=32"
+        psql ${BENCHMARK_CONNSTR} -c "ALTER DATABASE ${DB_NAME} SET maintenance_io_concurrency=32"
+      env:
+        BENCHMARK_CONNSTR: ${{ steps.set-up-connstr.outputs.connstr }}
+
    - name: Run user examples
      uses: ./.github/actions/run-python-test-set
      with:
@@ -565,7 +580,7 @@ jobs:
        BENCHMARK_CONNSTR: ${{ steps.set-up-connstr.outputs.connstr }}

    - name: Create Allure report
-      if: ${{ !cancelled() }}
+      if: success() || failure()
      uses: ./.github/actions/allure-report
      with:
        action: generate
--- a/.github/workflows/build_and_test.yml
+++ b/.github/workflows/build_and_test.yml
@@ -13,7 +13,7 @@ defaults:

 concurrency:
  # Allow only one workflow per any non-`main` branch.
-  group: ${{ github.workflow }}-${{ github.ref_name }}-${{ github.ref_name == 'main' && github.sha || 'anysha' }}
+  group: ${{ github.workflow }}-${{ github.ref }}-${{ github.ref == 'refs/heads/main' && github.sha || 'anysha' }}
  cancel-in-progress: true

 env:
@@ -335,10 +335,6 @@ jobs:
          real_s3_region: us-west-2
          real_s3_access_key_id: "${{ secrets.AWS_ACCESS_KEY_ID_CI_TESTS_S3 }}"
          real_s3_secret_access_key: "${{ secrets.AWS_SECRET_ACCESS_KEY_CI_TESTS_S3 }}"
-          rerun_flaky: true
-        env:
-          TEST_RESULT_CONNSTR: ${{ secrets.REGRESS_TEST_RESULT_CONNSTR }}
-          CHECK_ONDISK_DATA_COMPATIBILITY: nonempty

      - name: Merge and upload coverage data
        if: matrix.build_type == 'debug'
@@ -368,97 +364,49 @@ jobs:
          build_type: ${{ matrix.build_type }}
          test_selection: performance
          run_in_parallel: false
-          save_perf_report: ${{ github.ref_name == 'main' }}
+          save_perf_report: ${{ github.ref == 'refs/heads/main' }}
        env:
          VIP_VAP_ACCESS_TOKEN: "${{ secrets.VIP_VAP_ACCESS_TOKEN }}"
          PERF_TEST_RESULT_CONNSTR: "${{ secrets.PERF_TEST_RESULT_CONNSTR }}"
      # XXX: no coverage data handling here, since benchmarks are run on release builds,
      # while coverage is currently collected for the debug ones

-  create-test-report:
+  merge-allure-report:
    runs-on: [ self-hosted, gen3, small ]
    container:
      image: 369495373322.dkr.ecr.eu-central-1.amazonaws.com/rust:pinned
      options: --init
    needs: [ regress-tests, benchmarks ]
    if: ${{ !cancelled() }}
-
+    strategy:
+      fail-fast: false
+      matrix:
+        build_type: [ debug, release ]
    steps:
-      - uses: actions/checkout@v3
+      - name: Checkout
+        uses: actions/checkout@v3
+        with:
+          submodules: false

-      - name: Create Allure report (debug)
-        if: ${{ !cancelled() }}
-        id: create-allure-report-debug
+      - name: Create Allure report
+        id: create-allure-report
        uses: ./.github/actions/allure-report
        with:
          action: generate
-          build_type: debug
-
-      - name: Create Allure report (release)
-        if: ${{ !cancelled() }}
-        id: create-allure-report-release
-        uses: ./.github/actions/allure-report
-        with:
-          action: generate
-          build_type: release
-
-      - uses: actions/github-script@v6
-        if: >
-          !cancelled() &&
-          github.event_name == 'pull_request' && (
-            steps.create-allure-report-debug.outputs.report-url ||
-            steps.create-allure-report-release.outputs.report-url
-          )
-        with:
-          # Retry script for 5XX server errors: https://github.com/actions/github-script#retries
-          retries: 5
-          script: |
-            const reports = [{
-              buildType: "debug",
-              reportUrl: "${{ steps.create-allure-report-debug.outputs.report-url }}",
-              jsonUrl:   "${{ steps.create-allure-report-debug.outputs.report-json-url }}",
-            }, {
-              buildType: "release",
-              reportUrl: "${{ steps.create-allure-report-release.outputs.report-url }}",
-              jsonUrl:   "${{ steps.create-allure-report-release.outputs.report-json-url }}",
-            }]
-
-            const script = require("./scripts/pr-comment-test-report.js")
-            await script({
-              github,
-              context,
-              fetch,
-              reports,
-            })
+          build_type: ${{ matrix.build_type }}

      - name: Store Allure test stat in the DB
-        if: >
-          !cancelled() && (
-            steps.create-allure-report-debug.outputs.report-url ||
-            steps.create-allure-report-release.outputs.report-url
-          )
+        if: ${{ steps.create-allure-report.outputs.report-url }}
        env:
+          BUILD_TYPE: ${{ matrix.build_type }}
          SHA: ${{ github.event.pull_request.head.sha || github.sha }}
-          REPORT_JSON_URL_DEBUG: ${{ steps.create-allure-report-debug.outputs.report-json-url }}
-          REPORT_JSON_URL_RELEASE: ${{ steps.create-allure-report-release.outputs.report-json-url }}
+          REPORT_URL: ${{ steps.create-allure-report.outputs.report-url }}
          TEST_RESULT_CONNSTR: ${{ secrets.REGRESS_TEST_RESULT_CONNSTR }}
        run: |
+          curl --fail --output suites.json ${REPORT_URL%/index.html}/data/suites.json
          ./scripts/pysync

-          for report_url in $REPORT_JSON_URL_DEBUG $REPORT_JSON_URL_RELEASE; do
-            if [ -z "$report_url" ]; then
-              continue
-            fi
-
-            if [[ "$report_url" == "$REPORT_JSON_URL_DEBUG" ]]; then
-              BUILD_TYPE=debug
-            else
-              BUILD_TYPE=release
-            fi
-
-            curl --fail --output suites.json "${report_url}"
-            DATABASE_URL="$TEST_RESULT_CONNSTR" poetry run python3 scripts/ingest_regress_test_result.py --revision ${SHA} --reference ${GITHUB_REF} --build-type ${BUILD_TYPE} --ingest suites.json
-          done
+          DATABASE_URL="$TEST_RESULT_CONNSTR" poetry run python3 scripts/ingest_regress_test_result.py --revision ${SHA} --reference ${GITHUB_REF} --build-type ${BUILD_TYPE} --ingest suites.json

  coverage-report:
    runs-on: [ self-hosted, gen3, small ]
@@ -950,16 +898,6 @@ jobs:
    needs: [ push-docker-hub, tag, regress-tests ]
    if: ( github.ref_name == 'main' || github.ref_name == 'release' ) && github.event_name != 'workflow_dispatch'
    steps:
-      - name: Fix git ownership
-        run: |
-          # Workaround for `fatal: detected dubious ownership in repository at ...`
-          #
-          # Use both ${{ github.workspace }} and ${GITHUB_WORKSPACE} because they're different on host and in containers
-          #   Ref https://github.com/actions/checkout/issues/785
-          #
-          git config --global --add safe.directory ${{ github.workspace }}
-          git config --global --add safe.directory ${GITHUB_WORKSPACE}
-
      - name: Checkout
        uses: actions/checkout@v3
        with:
@@ -1007,7 +945,7 @@ jobs:

            S3_KEY=$(aws s3api list-objects-v2 --bucket ${BUCKET} --prefix ${OLD_PREFIX} | jq -r '.Contents[].Key' | grep ${FILENAME} | sort --version-sort | tail -1 || true)
            if [ -z "${S3_KEY}" ]; then
-              echo >&2 "Neither s3://${BUCKET}/${OLD_PREFIX}/${FILENAME} nor its version from previous attempts exist"
+              echo 2>&1 "Neither s3://${BUCKET}/${OLD_PREFIX}/${FILENAME} nor its version from previous attempts exist"
              exit 1
            fi

--- a/.github/workflows/neon_extra_builds.yml
+++ b/.github/workflows/neon_extra_builds.yml
@@ -12,7 +12,7 @@ defaults:

 concurrency:
  # Allow only one workflow per any non-`main` branch.
-  group: ${{ github.workflow }}-${{ github.ref_name }}-${{ github.ref_name == 'main' && github.sha || 'anysha' }}
+  group: ${{ github.workflow }}-${{ github.ref }}-${{ github.ref == 'refs/heads/main' && github.sha || 'anysha' }}
  cancel-in-progress: true

 env:
@@ -53,14 +53,14 @@ jobs:
        uses: actions/cache@v3
        with:
          path: pg_install/v14
-          key: v1-${{ runner.os }}-${{ env.BUILD_TYPE }}-pg-${{ steps.pg_v14_rev.outputs.pg_rev }}-${{ hashFiles('Makefile') }}
+          key: v1-${{ runner.os }}-${{ matrix.build_type }}-pg-${{ steps.pg_v14_rev.outputs.pg_rev }}-${{ hashFiles('Makefile') }}

      - name: Cache postgres v15 build
        id: cache_pg_15
        uses: actions/cache@v3
        with:
          path: pg_install/v15
-          key: v1-${{ runner.os }}-${{ env.BUILD_TYPE }}-pg-${{ steps.pg_v15_rev.outputs.pg_rev }}-${{ hashFiles('Makefile') }}
+          key: v1-${{ runner.os }}-${{ matrix.build_type }}-pg-${{ steps.pg_v15_rev.outputs.pg_rev }}-${{ hashFiles('Makefile') }}

      - name: Set extra env for macOS
        run: |
--- a/.github/workflows/pg_clients.yml
+++ b/.github/workflows/pg_clients.yml
@@ -14,7 +14,7 @@ on:

 concurrency:
  # Allow only one workflow per any non-`main` branch.
-  group: ${{ github.workflow }}-${{ github.ref_name }}-${{ github.ref_name == 'main' && github.sha || 'anysha' }}
+  group: ${{ github.workflow }}-${{ github.ref }}-${{ github.ref == 'refs/heads/main' && github.sha || 'anysha' }}
  cancel-in-progress: true

 jobs:
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -841,18 +841,6 @@ dependencies = [
 "unicode-width",
 ]

-[[package]]
-name = "compute_api"
-version = "0.1.0"
-dependencies = [
- "anyhow",
- "chrono",
- "serde",
- "serde_json",
- "serde_with",
- "workspace_hack",
-]
-
 [[package]]
 name = "compute_tools"
 version = "0.1.0"
@@ -860,7 +848,6 @@ dependencies = [
 "anyhow",
 "chrono",
 "clap 4.1.4",
- "compute_api",
 "futures",
 "hyper",
 "notify",
@@ -879,7 +866,6 @@ dependencies = [
 "tracing-subscriber",
 "tracing-utils",
 "url",
- "utils",
 "workspace_hack",
 ]

@@ -1572,9 +1558,9 @@ checksum = "d2fabcfbdc87f4758337ca535fb41a6d701b65693ce38287d856d1674551ec9b"

 [[package]]
 name = "h2"
-version = "0.3.17"
+version = "0.3.15"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "66b91535aa35fea1523ad1b86cb6b53c28e0ae566ba4a460f4457e936cad7c6f"
+checksum = "5f9f29bc9dda355256b2916cf526ab02ce0aeaaaf2bad60d65ef3f12f11dd0f4"
 dependencies = [
 "bytes",
 "fnv",
@@ -2488,7 +2474,6 @@ dependencies = [
 "strum",
 "strum_macros",
 "svg_fmt",
- "sync_wrapper",
 "tempfile",
 "tenant_size_model",
 "thiserror",
@@ -2517,8 +2502,6 @@ dependencies = [
 "serde",
 "serde_json",
 "serde_with",
- "strum",
- "strum_macros",
 "utils",
 "workspace_hack",
 ]
@@ -2965,7 +2948,6 @@ dependencies = [
 "tokio-postgres",
 "tokio-postgres-rustls",
 "tokio-rustls",
- "tokio-util",
 "tracing",
 "tracing-opentelemetry",
 "tracing-subscriber",
@@ -3371,7 +3353,6 @@ dependencies = [
 "tempfile",
 "thiserror",
 "tokio",
- "tokio-io-timeout",
 "tokio-postgres",
 "toml_edit",
 "tracing",
@@ -3991,16 +3972,6 @@ dependencies = [
 "time-core",
 ]

-[[package]]
-name = "timeline_data_path"
-version = "0.1.0"
-dependencies = [
- "thiserror",
- "tokio",
- "utils",
- "workspace_hack",
-]
-
 [[package]]
 name = "tinytemplate"
 version = "1.2.1"
@@ -4574,7 +4545,6 @@ dependencies = [
 "byteorder",
 "bytes",
 "criterion",
- "either",
 "futures",
 "heapless",
 "hex",
@@ -4586,7 +4556,6 @@ dependencies = [
 "once_cell",
 "pin-project-lite",
 "rand",
- "regex",
 "routerify",
 "sentry",
 "serde",
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -101,7 +101,6 @@ test-context = "0.1"
 thiserror = "1.0"
 tls-listener = { version = "0.6", features = ["rustls", "hyper-h1"] }
 tokio = { version = "1.17", features = ["macros"] }
-tokio-io-timeout = "1.2.0"
 tokio-postgres-rustls = "0.9.0"
 tokio-rustls = "0.23"
 tokio-stream = "0.1"
@@ -133,7 +132,6 @@ tokio-tar = { git = "https://github.com/neondatabase/tokio-tar.git", rev="404df6
 heapless = { default-features=false, features=[], git = "https://github.com/japaric/heapless.git", rev = "644653bf3b831c6bb4963be2de24804acf5e5001" } # upstream release pending

 ## Local libraries
-compute_api = { version = "0.1", path = "./libs/compute_api/" }
 consumption_metrics = { version = "0.1", path = "./libs/consumption_metrics/" }
 metrics = { version = "0.1", path = "./libs/metrics/" }
 pageserver_api = { version = "0.1", path = "./libs/pageserver_api/" }
--- a/Dockerfile.compute-node
+++ b/Dockerfile.compute-node
@@ -38,7 +38,6 @@ RUN cd postgres && \
    echo 'trusted = true' >> /usr/local/pgsql/share/extension/insert_username.control && \
    echo 'trusted = true' >> /usr/local/pgsql/share/extension/intagg.control && \
    echo 'trusted = true' >> /usr/local/pgsql/share/extension/moddatetime.control && \
-    echo 'trusted = true' >> /usr/local/pgsql/share/extension/pg_stat_statements.control && \
    echo 'trusted = true' >> /usr/local/pgsql/share/extension/pgrowlocks.control && \
    echo 'trusted = true' >> /usr/local/pgsql/share/extension/pgstattuple.control && \
    echo 'trusted = true' >> /usr/local/pgsql/share/extension/refint.control && \
@@ -60,7 +59,6 @@ RUN apt update && \

 # SFCGAL > 1.3 requires CGAL > 5.2, Bullseye's libcgal-dev is 5.2
 RUN wget https://gitlab.com/Oslandia/SFCGAL/-/archive/v1.3.10/SFCGAL-v1.3.10.tar.gz -O SFCGAL.tar.gz && \
-    echo "4e39b3b2adada6254a7bdba6d297bb28e1a9835a9f879b74f37e2dab70203232 SFCGAL.tar.gz" | sha256sum --check && \
    mkdir sfcgal-src && cd sfcgal-src && tar xvzf ../SFCGAL.tar.gz --strip-components=1 -C . && \
    cmake . && make -j $(getconf _NPROCESSORS_ONLN) && \
    DESTDIR=/sfcgal make install -j $(getconf _NPROCESSORS_ONLN) && \
@@ -69,7 +67,6 @@ RUN wget https://gitlab.com/Oslandia/SFCGAL/-/archive/v1.3.10/SFCGAL-v1.3.10.tar
 ENV PATH "/usr/local/pgsql/bin:$PATH"

 RUN wget https://download.osgeo.org/postgis/source/postgis-3.3.2.tar.gz -O postgis.tar.gz && \
-    echo "9a2a219da005a1730a39d1959a1c7cec619b1efb009b65be80ffc25bad299068 postgis.tar.gz" | sha256sum --check && \
    mkdir postgis-src && cd postgis-src && tar xvzf ../postgis.tar.gz --strip-components=1 -C . && \
    ./autogen.sh && \
    ./configure --with-sfcgal=/usr/local/bin/sfcgal-config && \
@@ -86,7 +83,6 @@ RUN wget https://download.osgeo.org/postgis/source/postgis-3.3.2.tar.gz -O postg
    echo 'trusted = true' >> /usr/local/pgsql/share/extension/address_standardizer_data_us.control

 RUN wget https://github.com/pgRouting/pgrouting/archive/v3.4.2.tar.gz -O pgrouting.tar.gz && \
-    echo "cac297c07d34460887c4f3b522b35c470138760fe358e351ad1db4edb6ee306e pgrouting.tar.gz" | sha256sum --check && \
    mkdir pgrouting-src && cd pgrouting-src && tar xvzf ../pgrouting.tar.gz --strip-components=1 -C . && \
    mkdir build && \
    cd build && \
@@ -107,7 +103,6 @@ RUN apt update && \
    apt install -y ninja-build python3-dev libncurses5 binutils clang

 RUN wget https://github.com/plv8/plv8/archive/refs/tags/v3.1.5.tar.gz -O plv8.tar.gz && \
-    echo "1e108d5df639e4c189e1c5bdfa2432a521c126ca89e7e5a969d46899ca7bf106 plv8.tar.gz" | sha256sum --check && \
    mkdir plv8-src && cd plv8-src && tar xvzf ../plv8.tar.gz --strip-components=1 -C . && \
    export PATH="/usr/local/pgsql/bin:$PATH" && \
    make DOCKER=1 -j $(getconf _NPROCESSORS_ONLN) install && \
@@ -129,13 +124,11 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 # packaged cmake is too old
 RUN wget https://github.com/Kitware/CMake/releases/download/v3.24.2/cmake-3.24.2-linux-x86_64.sh \
      -q -O /tmp/cmake-install.sh \
-      && echo "739d372726cb23129d57a539ce1432453448816e345e1545f6127296926b6754 /tmp/cmake-install.sh" | sha256sum --check \
      && chmod u+x /tmp/cmake-install.sh \
      && /tmp/cmake-install.sh --skip-license --prefix=/usr/local/ \
      && rm /tmp/cmake-install.sh

 RUN wget https://github.com/uber/h3/archive/refs/tags/v4.1.0.tar.gz -O h3.tar.gz && \
-    echo "ec99f1f5974846bde64f4513cf8d2ea1b8d172d2218ab41803bf6a63532272bc h3.tar.gz" | sha256sum --check && \
    mkdir h3-src && cd h3-src && tar xvzf ../h3.tar.gz --strip-components=1 -C . && \
    mkdir build && cd build && \
    cmake .. -DCMAKE_BUILD_TYPE=Release && \
@@ -145,7 +138,6 @@ RUN wget https://github.com/uber/h3/archive/refs/tags/v4.1.0.tar.gz -O h3.tar.gz
    rm -rf build

 RUN wget https://github.com/zachasme/h3-pg/archive/refs/tags/v4.1.2.tar.gz -O h3-pg.tar.gz && \
-    echo "c135aa45999b2ad1326d2537c1cadef96d52660838e4ca371706c08fdea1a956 h3-pg.tar.gz" | sha256sum --check && \
    mkdir h3-pg-src && cd h3-pg-src && tar xvzf ../h3-pg.tar.gz --strip-components=1 -C . && \
    export PATH="/usr/local/pgsql/bin:$PATH" && \
    make -j $(getconf _NPROCESSORS_ONLN) && \
@@ -163,7 +155,6 @@ FROM build-deps AS unit-pg-build
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/

 RUN wget https://github.com/df7cb/postgresql-unit/archive/refs/tags/7.7.tar.gz -O postgresql-unit.tar.gz && \
-    echo "411d05beeb97e5a4abf17572bfcfbb5a68d98d1018918feff995f6ee3bb03e79 postgresql-unit.tar.gz" | sha256sum --check && \
    mkdir postgresql-unit-src && cd postgresql-unit-src && tar xvzf ../postgresql-unit.tar.gz --strip-components=1 -C . && \
    make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
    make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
@@ -184,7 +175,6 @@ FROM build-deps AS vector-pg-build
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/

 RUN wget https://github.com/pgvector/pgvector/archive/refs/tags/v0.4.0.tar.gz -O pgvector.tar.gz && \
-    echo "b76cf84ddad452cc880a6c8c661d137ddd8679c000a16332f4f03ecf6e10bcc8 pgvector.tar.gz" | sha256sum --check && \
    mkdir pgvector-src && cd pgvector-src && tar xvzf ../pgvector.tar.gz --strip-components=1 -C . && \
    make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
    make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
@@ -201,7 +191,6 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/

 # 9742dab1b2f297ad3811120db7b21451bca2d3c9 made on 13/11/2021
 RUN wget https://github.com/michelp/pgjwt/archive/9742dab1b2f297ad3811120db7b21451bca2d3c9.tar.gz -O pgjwt.tar.gz && \
-    echo "cfdefb15007286f67d3d45510f04a6a7a495004be5b3aecb12cda667e774203f pgjwt.tar.gz" | sha256sum --check && \
    mkdir pgjwt-src && cd pgjwt-src && tar xvzf ../pgjwt.tar.gz --strip-components=1 -C . && \
    make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
    echo 'trusted = true' >> /usr/local/pgsql/share/extension/pgjwt.control
@@ -216,7 +205,6 @@ FROM build-deps AS hypopg-pg-build
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/

 RUN wget https://github.com/HypoPG/hypopg/archive/refs/tags/1.3.1.tar.gz -O hypopg.tar.gz && \
-    echo "e7f01ee0259dc1713f318a108f987663d60f3041948c2ada57a94b469565ca8e hypopg.tar.gz" | sha256sum --check && \
    mkdir hypopg-src && cd hypopg-src && tar xvzf ../hypopg.tar.gz --strip-components=1 -C . && \
    make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
    make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
@@ -232,7 +220,6 @@ FROM build-deps AS pg-hashids-pg-build
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/

 RUN wget https://github.com/iCyberon/pg_hashids/archive/refs/tags/v1.2.1.tar.gz -O pg_hashids.tar.gz && \
-    echo "74576b992d9277c92196dd8d816baa2cc2d8046fe102f3dcd7f3c3febed6822a pg_hashids.tar.gz" | sha256sum --check && \
    mkdir pg_hashids-src && cd pg_hashids-src && tar xvzf ../pg_hashids.tar.gz --strip-components=1 -C . && \
    make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config USE_PGXS=1 && \
    make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config USE_PGXS=1 && \
@@ -248,7 +235,6 @@ FROM build-deps AS rum-pg-build
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/

 RUN wget https://github.com/postgrespro/rum/archive/refs/tags/1.3.13.tar.gz -O rum.tar.gz && \
-    echo "6ab370532c965568df6210bd844ac6ba649f53055e48243525b0b7e5c4d69a7d rum.tar.gz" | sha256sum --check && \
    mkdir rum-src && cd rum-src && tar xvzf ../rum.tar.gz --strip-components=1 -C . && \
    make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config USE_PGXS=1 && \
    make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config USE_PGXS=1 && \
@@ -264,28 +250,11 @@ FROM build-deps AS pgtap-pg-build
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/

 RUN wget https://github.com/theory/pgtap/archive/refs/tags/v1.2.0.tar.gz -O pgtap.tar.gz && \
-    echo "9c7c3de67ea41638e14f06da5da57bac6f5bd03fea05c165a0ec862205a5c052 pgtap.tar.gz" | sha256sum --check && \
    mkdir pgtap-src && cd pgtap-src && tar xvzf ../pgtap.tar.gz --strip-components=1 -C . && \
    make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
    make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
    echo 'trusted = true' >> /usr/local/pgsql/share/extension/pgtap.control

-#########################################################################################
-#
-# Layer "ip4r-pg-build"
-# compile ip4r extension
-#
-#########################################################################################
-FROM build-deps AS ip4r-pg-build
-COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
-
-RUN wget https://github.com/RhodiumToad/ip4r/archive/refs/tags/2.4.1.tar.gz -O ip4r.tar.gz && \
-    echo "78b9f0c1ae45c22182768fe892a32d533c82281035e10914111400bf6301c726 ip4r.tar.gz" | sha256sum --check && \
-    mkdir ip4r-src && cd ip4r-src && tar xvzf ../ip4r.tar.gz --strip-components=1 -C . && \
-    make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
-    make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
-    echo 'trusted = true' >> /usr/local/pgsql/share/extension/ip4r.control
-
 #########################################################################################
 #
 # Layer "prefix-pg-build"
@@ -296,7 +265,6 @@ FROM build-deps AS prefix-pg-build
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/

 RUN wget https://github.com/dimitri/prefix/archive/refs/tags/v1.2.9.tar.gz -O prefix.tar.gz && \
-    echo "38d30a08d0241a8bbb8e1eb8f0152b385051665a8e621c8899e7c5068f8b511e prefix.tar.gz" | sha256sum --check && \
    mkdir prefix-src && cd prefix-src && tar xvzf ../prefix.tar.gz --strip-components=1 -C . && \
    make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
    make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
@@ -312,7 +280,6 @@ FROM build-deps AS hll-pg-build
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/

 RUN wget https://github.com/citusdata/postgresql-hll/archive/refs/tags/v2.17.tar.gz -O hll.tar.gz && \
-    echo "9a18288e884f197196b0d29b9f178ba595b0dfc21fbf7a8699380e77fa04c1e9 hll.tar.gz" | sha256sum --check && \
    mkdir hll-src && cd hll-src && tar xvzf ../hll.tar.gz --strip-components=1 -C . && \
    make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
    make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
@@ -328,68 +295,13 @@ FROM build-deps AS plpgsql-check-pg-build
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/

 RUN wget https://github.com/okbob/plpgsql_check/archive/refs/tags/v2.3.2.tar.gz -O plpgsql_check.tar.gz && \
-    echo "9d81167c4bbeb74eebf7d60147b21961506161addc2aee537f95ad8efeae427b plpgsql_check.tar.gz" | sha256sum --check && \
    mkdir plpgsql_check-src && cd plpgsql_check-src && tar xvzf ../plpgsql_check.tar.gz --strip-components=1 -C . && \
    make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config USE_PGXS=1 && \
    make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config USE_PGXS=1 && \
    echo 'trusted = true' >> /usr/local/pgsql/share/extension/plpgsql_check.control

 #########################################################################################
-#
-# Layer "timescaledb-pg-build"
-# compile timescaledb extension
-#
-#########################################################################################
-FROM build-deps AS timescaledb-pg-build
-COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
-
-ENV PATH "/usr/local/pgsql/bin:$PATH"
-
-RUN apt-get update && \
-    apt-get install -y cmake && \
-    wget https://github.com/timescale/timescaledb/archive/refs/tags/2.10.1.tar.gz -O timescaledb.tar.gz && \
-    echo "6fca72a6ed0f6d32d2b3523951ede73dc5f9b0077b38450a029a5f411fdb8c73 timescaledb.tar.gz" | sha256sum --check && \
-    mkdir timescaledb-src && cd timescaledb-src && tar xvzf ../timescaledb.tar.gz --strip-components=1 -C . && \
-    ./bootstrap -DSEND_TELEMETRY_DEFAULT:BOOL=OFF -DUSE_TELEMETRY:BOOL=OFF -DAPACHE_ONLY:BOOL=ON && \
-    cd build && \
-    make -j $(getconf _NPROCESSORS_ONLN) && \
-    make install -j $(getconf _NPROCESSORS_ONLN) && \
-    echo "trusted = true" >> /usr/local/pgsql/share/extension/timescaledb.control
-
-#########################################################################################
-#
-# Layer "pg-hint-plan-pg-build"
-# compile pg_hint_plan extension
-#
-#########################################################################################
-FROM build-deps AS pg-hint-plan-pg-build
-COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
-
-ARG PG_VERSION
-ENV PATH "/usr/local/pgsql/bin:$PATH"
-
-RUN case "${PG_VERSION}" in \
-      "v14") \
-        export PG_HINT_PLAN_VERSION=14_1_4_1 \
-        export PG_HINT_PLAN_CHECKSUM=c3501becf70ead27f70626bce80ea401ceac6a77e2083ee5f3ff1f1444ec1ad1 \
-        ;; \
-      "v15") \
-        export PG_HINT_PLAN_VERSION=15_1_5_0 \
-        export PG_HINT_PLAN_CHECKSUM=564cbbf4820973ffece63fbf76e3c0af62c4ab23543142c7caaa682bc48918be \
-        ;; \
-      *) \
-        echo "Export the valid PG_HINT_PLAN_VERSION variable" && exit 1 \
-        ;; \
-    esac && \
-    wget https://github.com/ossc-db/pg_hint_plan/archive/refs/tags/REL${PG_HINT_PLAN_VERSION}.tar.gz -O pg_hint_plan.tar.gz && \
-    echo "${PG_HINT_PLAN_CHECKSUM} pg_hint_plan.tar.gz" | sha256sum --check && \
-    mkdir pg_hint_plan-src && cd pg_hint_plan-src && tar xvzf ../pg_hint_plan.tar.gz --strip-components=1 -C . && \
-    make -j $(getconf _NPROCESSORS_ONLN) && \
-    make install -j $(getconf _NPROCESSORS_ONLN) && \
-    echo "trusted = true" >> /usr/local/pgsql/share/extension/pg_hint_plan.control
-
-#########################################################################################
-#
+# 
 # Layer "rust extensions"
 # This layer is used to build `pgx` deps
 #
@@ -417,7 +329,7 @@ RUN curl -sSO https://static.rust-lang.org/rustup/dist/$(uname -m)-unknown-linux
 USER root

 #########################################################################################
-#
+# 
 # Layer "pg-jsonschema-pg-build"
 # Compile "pg_jsonschema" extension
 #
@@ -425,17 +337,15 @@ USER root

 FROM rust-extensions-build AS pg-jsonschema-pg-build

-# caeab60d70b2fd3ae421ec66466a3abbb37b7ee6 made on 06/03/2023
-# there is no release tag yet, but we need it due to the superuser fix in the control file, switch to git tag after release >= 0.1.5
+# there is no release tag yet, but we need it due to the superuser fix in the control file
 RUN wget https://github.com/supabase/pg_jsonschema/archive/caeab60d70b2fd3ae421ec66466a3abbb37b7ee6.tar.gz -O pg_jsonschema.tar.gz && \
-    echo "54129ce2e7ee7a585648dbb4cef6d73f795d94fe72f248ac01119992518469a4 pg_jsonschema.tar.gz" | sha256sum --check && \
    mkdir pg_jsonschema-src && cd pg_jsonschema-src && tar xvzf ../pg_jsonschema.tar.gz --strip-components=1 -C . && \
    sed -i 's/pgx = "0.7.1"/pgx = { version = "0.7.3", features = [ "unsafe-postgres" ] }/g' Cargo.toml && \
    cargo pgx install --release && \
    echo "trusted = true" >> /usr/local/pgsql/share/extension/pg_jsonschema.control

 #########################################################################################
-#
+# 
 # Layer "pg-graphql-pg-build"
 # Compile "pg_graphql" extension
 #
@@ -443,13 +353,11 @@ RUN wget https://github.com/supabase/pg_jsonschema/archive/caeab60d70b2fd3ae421e

 FROM rust-extensions-build AS pg-graphql-pg-build

-# b4988843647450a153439be367168ed09971af85 made on 22/02/2023 (from remove-pgx-contrib-spiext branch)
 # Currently pgx version bump to >= 0.7.2  causes "call to unsafe function" compliation errors in
 # pgx-contrib-spiext. There is a branch that removes that dependency, so use it. It is on the
 # same 1.1 version we've used before.
-RUN wget https://github.com/yrashk/pg_graphql/archive/b4988843647450a153439be367168ed09971af85.tar.gz -O pg_graphql.tar.gz && \
-    echo "0c7b0e746441b2ec24187d0e03555faf935c2159e2839bddd14df6dafbc8c9bd pg_graphql.tar.gz" | sha256sum --check && \
-    mkdir pg_graphql-src && cd pg_graphql-src && tar xvzf ../pg_graphql.tar.gz --strip-components=1 -C . && \
+RUN git clone -b remove-pgx-contrib-spiext --single-branch https://github.com/yrashk/pg_graphql && \
+    cd pg_graphql && \
    sed -i 's/pgx = "~0.7.1"/pgx = { version = "0.7.3", features = [ "unsafe-postgres" ] }/g' Cargo.toml && \
    sed -i 's/pgx-tests = "~0.7.1"/pgx-tests = "0.7.3"/g' Cargo.toml && \
    cargo pgx install --release && \
@@ -466,10 +374,8 @@ RUN wget https://github.com/yrashk/pg_graphql/archive/b4988843647450a153439be367

 FROM rust-extensions-build AS pg-tiktoken-pg-build

-# 801f84f08c6881c8aa30f405fafbf00eec386a72 made on 10/03/2023
-RUN wget https://github.com/kelvich/pg_tiktoken/archive/801f84f08c6881c8aa30f405fafbf00eec386a72.tar.gz -O pg_tiktoken.tar.gz && \
-    echo "52f60ac800993a49aa8c609961842b611b6b1949717b69ce2ec9117117e16e4a pg_tiktoken.tar.gz" | sha256sum --check && \
-    mkdir pg_tiktoken-src && cd pg_tiktoken-src && tar xvzf ../pg_tiktoken.tar.gz --strip-components=1 -C . && \
+RUN git clone --depth=1 --single-branch https://github.com/kelvich/pg_tiktoken && \
+    cd pg_tiktoken && \
    cargo pgx install --release && \
    echo "trusted = true" >> /usr/local/pgsql/share/extension/pg_tiktoken.control

@@ -495,12 +401,9 @@ COPY --from=hypopg-pg-build /usr/local/pgsql/ /usr/local/pgsql/
 COPY --from=pg-hashids-pg-build /usr/local/pgsql/ /usr/local/pgsql/
 COPY --from=rum-pg-build /usr/local/pgsql/ /usr/local/pgsql/
 COPY --from=pgtap-pg-build /usr/local/pgsql/ /usr/local/pgsql/
-COPY --from=ip4r-pg-build /usr/local/pgsql/ /usr/local/pgsql/
 COPY --from=prefix-pg-build /usr/local/pgsql/ /usr/local/pgsql/
 COPY --from=hll-pg-build /usr/local/pgsql/ /usr/local/pgsql/
 COPY --from=plpgsql-check-pg-build /usr/local/pgsql/ /usr/local/pgsql/
-COPY --from=timescaledb-pg-build /usr/local/pgsql/ /usr/local/pgsql/
-COPY --from=pg-hint-plan-pg-build /usr/local/pgsql/ /usr/local/pgsql/
 COPY pgxn/ pgxn/

 RUN make -j $(getconf _NPROCESSORS_ONLN) \
--- a/README.md
+++ b/README.md
@@ -40,8 +40,6 @@ pacman -S base-devel readline zlib libseccomp openssl clang \
 postgresql-libs cmake postgresql protobuf
 ```

-Building Neon requires 3.15+ version of `protoc` (protobuf-compiler). If your distribution provides an older version, you can install a newer version from [here](https://github.com/protocolbuffers/protobuf/releases).
-
 2. [Install Rust](https://www.rust-lang.org/tools/install)
 ```
 # recommended approach from https://www.rust-lang.org/tools/install
@@ -147,15 +145,15 @@ Created an initial timeline 'de200bd42b49cc1814412c7e592dd6e9' at Lsn 0/16B5A50
 Setting tenant 9ef87a5bf0d92544f6fafeeb3239695c as a default one

 # start postgres compute node
-> ./target/debug/neon_local endpoint start main
-Starting new endpoint main (PostgreSQL v14) on timeline de200bd42b49cc1814412c7e592dd6e9 ...
+> ./target/debug/neon_local pg start main
+Starting new postgres (v14) main on timeline de200bd42b49cc1814412c7e592dd6e9 ...
 Extracting base backup to create postgres instance: path=.neon/pgdatadirs/tenants/9ef87a5bf0d92544f6fafeeb3239695c/main port=55432
-Starting postgres at 'host=127.0.0.1 port=55432 user=cloud_admin dbname=postgres'
+Starting postgres node at 'host=127.0.0.1 port=55432 user=cloud_admin dbname=postgres'

 # check list of running postgres instances
-> ./target/debug/neon_local endpoint list
- ENDPOINT  ADDRESS          TIMELINE                          BRANCH NAME  LSN        STATUS
- main      127.0.0.1:55432  de200bd42b49cc1814412c7e592dd6e9  main         0/16B5BA8  running
+> ./target/debug/neon_local pg list
+ NODE  ADDRESS          TIMELINE                          BRANCH NAME  LSN        STATUS
+ main  127.0.0.1:55432  de200bd42b49cc1814412c7e592dd6e9  main         0/16B5BA8  running
 ```

 2. Now, it is possible to connect to postgres and run some queries:
@@ -184,14 +182,14 @@ Created timeline 'b3b863fa45fa9e57e615f9f2d944e601' at Lsn 0/16F9A00 for tenant:
 (L) ┗━ @0/16F9A00: migration_check [b3b863fa45fa9e57e615f9f2d944e601]

 # start postgres on that branch
-> ./target/debug/neon_local endpoint start migration_check --branch-name migration_check
-Starting new endpoint migration_check (PostgreSQL v14) on timeline b3b863fa45fa9e57e615f9f2d944e601 ...
+> ./target/debug/neon_local pg start migration_check --branch-name migration_check
+Starting new postgres migration_check on timeline b3b863fa45fa9e57e615f9f2d944e601 ...
 Extracting base backup to create postgres instance: path=.neon/pgdatadirs/tenants/9ef87a5bf0d92544f6fafeeb3239695c/migration_check port=55433
-Starting postgres at 'host=127.0.0.1 port=55433 user=cloud_admin dbname=postgres'
+Starting postgres node at 'host=127.0.0.1 port=55433 user=cloud_admin dbname=postgres'

 # check the new list of running postgres instances
-> ./target/debug/neon_local endpoint list
- ENDPOINT         ADDRESS          TIMELINE                          BRANCH NAME      LSN        STATUS
+> ./target/debug/neon_local pg list
+ NODE             ADDRESS          TIMELINE                          BRANCH NAME      LSN        STATUS
 main             127.0.0.1:55432  de200bd42b49cc1814412c7e592dd6e9  main             0/16F9A38  running
 migration_check  127.0.0.1:55433  b3b863fa45fa9e57e615f9f2d944e601  migration_check  0/16F9A70  running

--- a/compute_tools/Cargo.toml
+++ b/compute_tools/Cargo.toml
@@ -27,6 +27,4 @@ tracing-subscriber.workspace = true
 tracing-utils.workspace = true
 url.workspace = true

-compute_api.workspace = true
-utils.workspace = true
 workspace_hack.workspace = true
--- a/compute_tools/src/bin/compute_ctl.rs
+++ b/compute_tools/src/bin/compute_ctl.rs
@@ -34,24 +34,22 @@ use std::fs::File;
 use std::panic;
 use std::path::Path;
 use std::process::exit;
-use std::sync::{mpsc, Arc, Condvar, Mutex};
+use std::sync::{Arc, RwLock};
 use std::{thread, time::Duration};

 use anyhow::{Context, Result};
 use chrono::Utc;
 use clap::Arg;
 use tracing::{error, info};
-use url::Url;

-use compute_api::responses::ComputeStatus;
-
-use compute_tools::compute::{ComputeNode, ComputeState, ParsedSpec};
-use compute_tools::configurator::launch_configurator;
+use compute_tools::compute::{ComputeMetrics, ComputeNode, ComputeState, ComputeStatus};
 use compute_tools::http::api::launch_http_server;
 use compute_tools::logger::*;
 use compute_tools::monitor::launch_monitor;
 use compute_tools::params::*;
+use compute_tools::pg_helpers::*;
 use compute_tools::spec::*;
+use url::Url;

 fn main() -> Result<()> {
    init_tracing_and_logging(DEFAULT_LOG_LEVEL)?;
@@ -64,7 +62,7 @@ fn main() -> Result<()> {
    let connstr = matches
        .get_one::<String>("connstr")
        .expect("Postgres connection string is required");
-    let spec_json = matches.get_one::<String>("spec");
+    let spec = matches.get_one::<String>("spec");
    let spec_path = matches.get_one::<String>("spec-path");

    let compute_id = matches.get_one::<String>("compute-id");
@@ -73,84 +71,40 @@ fn main() -> Result<()> {
    // Try to use just 'postgres' if no path is provided
    let pgbin = matches.get_one::<String>("pgbin").unwrap();

-    let mut spec = None;
-    let mut live_config_allowed = false;
-    match spec_json {
+    let spec: ComputeSpec = match spec {
        // First, try to get cluster spec from the cli argument
-        Some(json) => {
-            spec = Some(serde_json::from_str(json)?);
-        }
+        Some(json) => serde_json::from_str(json)?,
        None => {
            // Second, try to read it from the file if path is provided
            if let Some(sp) = spec_path {
                let path = Path::new(sp);
                let file = File::open(path)?;
-                spec = Some(serde_json::from_reader(file)?);
+                serde_json::from_reader(file)?
            } else if let Some(id) = compute_id {
                if let Some(cp_base) = control_plane_uri {
-                    live_config_allowed = true;
-                    if let Ok(s) = get_spec_from_control_plane(cp_base, id) {
-                        spec = Some(s);
-                    }
+                    let cp_uri = format!("{cp_base}/management/api/v1/{id}/spec");
+                    let jwt: String = match std::env::var("NEON_CONSOLE_JWT") {
+                        Ok(v) => v,
+                        Err(_) => "".to_string(),
+                    };
+
+                    reqwest::blocking::Client::new()
+                        .get(cp_uri)
+                        .header("Authorization", jwt)
+                        .send()?
+                        .json()?
                } else {
-                    panic!("must specify both --control-plane-uri and --compute-id or none");
+                    panic!(
+                        "must specify --control-plane-uri \"{:#?}\" and --compute-id \"{:#?}\"",
+                        control_plane_uri, compute_id
+                    );
                }
            } else {
-                panic!(
-                    "compute spec should be provided by one of the following ways: \
-                    --spec OR --spec-path OR --control-plane-uri and --compute-id"
-                );
+                panic!("compute spec should be provided via --spec or --spec-path argument");
            }
        }
    };

-    let mut new_state = ComputeState::new();
-    let spec_set;
-    if let Some(spec) = spec {
-        let pspec = ParsedSpec::try_from(spec).map_err(|msg| anyhow::anyhow!(msg))?;
-        new_state.pspec = Some(pspec);
-        spec_set = true;
-    } else {
-        spec_set = false;
-    }
-    let compute_node = ComputeNode {
-        start_time: Utc::now(),
-        connstr: Url::parse(connstr).context("cannot parse connstr as a URL")?,
-        pgdata: pgdata.to_string(),
-        pgbin: pgbin.to_string(),
-        live_config_allowed,
-        state: Mutex::new(new_state),
-        state_changed: Condvar::new(),
-    };
-    let compute = Arc::new(compute_node);
-
-    // Launch http service first, so we were able to serve control-plane
-    // requests, while configuration is still in progress.
-    let _http_handle = launch_http_server(&compute).expect("cannot launch http endpoint thread");
-
-    if !spec_set {
-        // No spec provided, hang waiting for it.
-        info!("no compute spec provided, waiting");
-        let mut state = compute.state.lock().unwrap();
-        while state.status != ComputeStatus::ConfigurationPending {
-            state = compute.state_changed.wait(state).unwrap();
-
-            if state.status == ComputeStatus::ConfigurationPending {
-                info!("got spec, continue configuration");
-                // Spec is already set by the http server handler.
-                break;
-            }
-        }
-    }
-
-    // We got all we need, update the state.
-    let mut state = compute.state.lock().unwrap();
-    let pspec = state.pspec.as_ref().expect("spec must be set");
-    let startup_tracing_context = pspec.spec.startup_tracing_context.clone();
-    state.status = ComputeStatus::Init;
-    compute.state_changed.notify_all();
-    drop(state);
-
    // Extract OpenTelemetry context for the startup actions from the spec, and
    // attach it to the current tracing context.
    //
@@ -166,7 +120,7 @@ fn main() -> Result<()> {
    // postgres is configured and up-and-running, we exit this span. Any other
    // actions that are performed on incoming HTTP requests, for example, are
    // performed in separate spans.
-    let startup_context_guard = if let Some(ref carrier) = startup_tracing_context {
+    let startup_context_guard = if let Some(ref carrier) = spec.startup_tracing_context {
        use opentelemetry::propagation::TextMapPropagator;
        use opentelemetry::sdk::propagation::TraceContextPropagator;
        Some(TraceContextPropagator::new().extract(carrier).attach())
@@ -174,10 +128,42 @@ fn main() -> Result<()> {
        None
    };

-    // Launch remaining service threads
+    let pageserver_connstr = spec
+        .cluster
+        .settings
+        .find("neon.pageserver_connstring")
+        .expect("pageserver connstr should be provided");
+    let storage_auth_token = spec.storage_auth_token.clone();
+    let tenant = spec
+        .cluster
+        .settings
+        .find("neon.tenant_id")
+        .expect("tenant id should be provided");
+    let timeline = spec
+        .cluster
+        .settings
+        .find("neon.timeline_id")
+        .expect("tenant id should be provided");
+
+    let compute_state = ComputeNode {
+        start_time: Utc::now(),
+        connstr: Url::parse(connstr).context("cannot parse connstr as a URL")?,
+        pgdata: pgdata.to_string(),
+        pgbin: pgbin.to_string(),
+        spec,
+        tenant,
+        timeline,
+        pageserver_connstr,
+        storage_auth_token,
+        metrics: ComputeMetrics::default(),
+        state: RwLock::new(ComputeState::new()),
+    };
+    let compute = Arc::new(compute_state);
+
+    // Launch service threads first, so we were able to serve availability
+    // requests, while configuration is still in progress.
+    let _http_handle = launch_http_server(&compute).expect("cannot launch http endpoint thread");
    let _monitor_handle = launch_monitor(&compute).expect("cannot launch compute monitor thread");
-    let _configurator_handle =
-        launch_configurator(&compute).expect("cannot launch configurator thread");

    // Start Postgres
    let mut delay_exit = false;
@@ -186,7 +172,7 @@ fn main() -> Result<()> {
        Ok(pg) => Some(pg),
        Err(err) => {
            error!("could not start the compute node: {:?}", err);
-            let mut state = compute.state.lock().unwrap();
+            let mut state = compute.state.write().unwrap();
            state.error = Some(format!("{:?}", err));
            state.status = ComputeStatus::Failed;
            drop(state);
@@ -219,25 +205,10 @@ fn main() -> Result<()> {
        thread::sleep(Duration::from_secs(30));
    }

-    // Shutdown trace pipeline gracefully, so that it has a chance to send any
-    // pending traces before we exit. Shutting down OTEL tracing provider may
-    // hang for quite some time, see, for example:
-    // - https://github.com/open-telemetry/opentelemetry-rust/issues/868
-    // - and our problems with staging https://github.com/neondatabase/cloud/issues/3707#issuecomment-1493983636
-    //
-    // Yet, we want computes to shut down fast enough, as we may need a new one
-    // for the same timeline ASAP. So wait no longer than 2s for the shutdown to
-    // complete, then just error out and exit the main thread.
    info!("shutting down tracing");
-    let (sender, receiver) = mpsc::channel();
-    let _ = thread::spawn(move || {
-        tracing_utils::shutdown_tracing();
-        sender.send(()).ok()
-    });
-    let shutdown_res = receiver.recv_timeout(Duration::from_millis(2000));
-    if shutdown_res.is_err() {
-        error!("timed out while shutting down tracing, exiting anyway");
-    }
+    // Shutdown trace pipeline gracefully, so that it has a chance to send any
+    // pending traces before we exit.
+    tracing_utils::shutdown_tracing();

    info!("shutting down");
    exit(exit_code.unwrap_or(1))
@@ -291,7 +262,7 @@ fn cli() -> clap::Command {
            Arg::new("control-plane-uri")
                .short('p')
                .long("control-plane-uri")
-                .value_name("CONTROL_PLANE_API_BASE_URI"),
+                .value_name("CONTROL_PLANE"),
        )
 }

--- a/compute_tools/src/compute.rs
+++ b/compute_tools/src/compute.rs
@@ -19,18 +19,15 @@ use std::os::unix::fs::PermissionsExt;
 use std::path::Path;
 use std::process::{Command, Stdio};
 use std::str::FromStr;
-use std::sync::{Condvar, Mutex};
+use std::sync::atomic::{AtomicU64, Ordering};
+use std::sync::RwLock;

 use anyhow::{Context, Result};
 use chrono::{DateTime, Utc};
 use postgres::{Client, NoTls};
+use serde::{Serialize, Serializer};
 use tokio_postgres;
 use tracing::{info, instrument, warn};
-use utils::id::{TenantId, TimelineId};
-use utils::lsn::Lsn;
-
-use compute_api::responses::{ComputeMetrics, ComputeStatus};
-use compute_api::spec::ComputeSpec;

 use crate::checker::create_writability_check_data;
 use crate::config;
@@ -44,45 +41,41 @@ pub struct ComputeNode {
    pub connstr: url::Url,
    pub pgdata: String,
    pub pgbin: String,
-    /// We should only allow live re- / configuration of the compute node if
-    /// it uses 'pull model', i.e. it can go to control-plane and fetch
-    /// the latest configuration. Otherwise, there could be a case:
-    /// - we start compute with some spec provided as argument
-    /// - we push new spec and it does reconfiguration
-    /// - but then something happens and compute pod / VM is destroyed,
-    ///   so k8s controller starts it again with the **old** spec
-    /// and the same for empty computes:
-    /// - we started compute without any spec
-    /// - we push spec and it does configuration
-    /// - but then it is restarted without any spec again
-    pub live_config_allowed: bool,
-    /// Volatile part of the `ComputeNode`, which should be used under `Mutex`.
-    /// To allow HTTP API server to serving status requests, while configuration
-    /// is in progress, lock should be held only for short periods of time to do
-    /// read/write, not the whole configuration process.
-    pub state: Mutex<ComputeState>,
-    /// `Condvar` to allow notifying waiters about state changes.
-    pub state_changed: Condvar,
+    pub spec: ComputeSpec,
+    pub tenant: String,
+    pub timeline: String,
+    pub pageserver_connstr: String,
+    pub storage_auth_token: Option<String>,
+    pub metrics: ComputeMetrics,
+    /// Volatile part of the `ComputeNode` so should be used under `RwLock`
+    /// to allow HTTP API server to serve status requests, while configuration
+    /// is in progress.
+    pub state: RwLock<ComputeState>,
 }

-#[derive(Clone, Debug)]
+fn rfc3339_serialize<S>(x: &DateTime<Utc>, s: S) -> Result<S::Ok, S::Error>
+where
+    S: Serializer,
+{
+    x.to_rfc3339().serialize(s)
+}
+
+#[derive(Serialize)]
+#[serde(rename_all = "snake_case")]
 pub struct ComputeState {
    pub status: ComputeStatus,
    /// Timestamp of the last Postgres activity
+    #[serde(serialize_with = "rfc3339_serialize")]
    pub last_active: DateTime<Utc>,
    pub error: Option<String>,
-    pub pspec: Option<ParsedSpec>,
-    pub metrics: ComputeMetrics,
 }

 impl ComputeState {
    pub fn new() -> Self {
        Self {
-            status: ComputeStatus::Empty,
+            status: ComputeStatus::Init,
            last_active: Utc::now(),
            error: None,
-            pspec: None,
-            metrics: ComputeMetrics::default(),
        }
    }
 }
@@ -93,58 +86,29 @@ impl Default for ComputeState {
    }
 }

-#[derive(Clone, Debug)]
-pub struct ParsedSpec {
-    pub spec: ComputeSpec,
-    pub tenant_id: TenantId,
-    pub timeline_id: TimelineId,
-    pub pageserver_connstr: String,
-    pub storage_auth_token: Option<String>,
+#[derive(Serialize, Clone, Copy, PartialEq, Eq)]
+#[serde(rename_all = "snake_case")]
+pub enum ComputeStatus {
+    Init,
+    Running,
+    Failed,
 }

-impl TryFrom<ComputeSpec> for ParsedSpec {
-    type Error = String;
-    fn try_from(spec: ComputeSpec) -> Result<Self, String> {
-        let pageserver_connstr = spec
-            .cluster
-            .settings
-            .find("neon.pageserver_connstring")
-            .ok_or("pageserver connstr should be provided")?;
-        let storage_auth_token = spec.storage_auth_token.clone();
-        let tenant_id: TenantId = spec
-            .cluster
-            .settings
-            .find("neon.tenant_id")
-            .ok_or("tenant id should be provided")
-            .map(|s| TenantId::from_str(&s))?
-            .or(Err("invalid tenant id"))?;
-        let timeline_id: TimelineId = spec
-            .cluster
-            .settings
-            .find("neon.timeline_id")
-            .ok_or("timeline id should be provided")
-            .map(|s| TimelineId::from_str(&s))?
-            .or(Err("invalid timeline id"))?;
-
-        Ok(ParsedSpec {
-            spec,
-            pageserver_connstr,
-            storage_auth_token,
-            tenant_id,
-            timeline_id,
-        })
-    }
+#[derive(Default, Serialize)]
+pub struct ComputeMetrics {
+    pub sync_safekeepers_ms: AtomicU64,
+    pub basebackup_ms: AtomicU64,
+    pub config_ms: AtomicU64,
+    pub total_startup_ms: AtomicU64,
 }

 impl ComputeNode {
    pub fn set_status(&self, status: ComputeStatus) {
-        let mut state = self.state.lock().unwrap();
-        state.status = status;
-        self.state_changed.notify_all();
+        self.state.write().unwrap().status = status;
    }

    pub fn get_status(&self) -> ComputeStatus {
-        self.state.lock().unwrap().status
+        self.state.read().unwrap().status
    }

    // Remove `pgdata` directory and create it again with right permissions.
@@ -160,16 +124,15 @@ impl ComputeNode {

    // Get basebackup from the libpq connection to pageserver using `connstr` and
    // unarchive it to `pgdata` directory overriding all its previous content.
-    #[instrument(skip(self, compute_state))]
-    fn get_basebackup(&self, compute_state: &ComputeState, lsn: Lsn) -> Result<()> {
-        let spec = compute_state.pspec.as_ref().expect("spec must be set");
+    #[instrument(skip(self))]
+    fn get_basebackup(&self, lsn: &str) -> Result<()> {
        let start_time = Utc::now();

-        let mut config = postgres::Config::from_str(&spec.pageserver_connstr)?;
+        let mut config = postgres::Config::from_str(&self.pageserver_connstr)?;

        // Use the storage auth token from the config file, if given.
        // Note: this overrides any password set in the connection string.
-        if let Some(storage_auth_token) = &spec.storage_auth_token {
+        if let Some(storage_auth_token) = &self.storage_auth_token {
            info!("Got storage auth token from spec file");
            config.password(storage_auth_token);
        } else {
@@ -178,8 +141,8 @@ impl ComputeNode {

        let mut client = config.connect(NoTls)?;
        let basebackup_cmd = match lsn {
-            Lsn(0) => format!("basebackup {} {}", spec.tenant_id, spec.timeline_id), // First start of the compute
-            _ => format!("basebackup {} {} {}", spec.tenant_id, spec.timeline_id, lsn),
+            "0/0" => format!("basebackup {} {}", &self.tenant, &self.timeline), // First start of the compute
+            _ => format!("basebackup {} {} {}", &self.tenant, &self.timeline, lsn),
        };
        let copyreader = client.copy_out(basebackup_cmd.as_str())?;

@@ -192,24 +155,28 @@ impl ComputeNode {
        ar.set_ignore_zeros(true);
        ar.unpack(&self.pgdata)?;

-        self.state.lock().unwrap().metrics.basebackup_ms = Utc::now()
-            .signed_duration_since(start_time)
-            .to_std()
-            .unwrap()
-            .as_millis() as u64;
+        self.metrics.basebackup_ms.store(
+            Utc::now()
+                .signed_duration_since(start_time)
+                .to_std()
+                .unwrap()
+                .as_millis() as u64,
+            Ordering::Relaxed,
+        );
+
        Ok(())
    }

    // Run `postgres` in a special mode with `--sync-safekeepers` argument
    // and return the reported LSN back to the caller.
-    #[instrument(skip(self, storage_auth_token))]
-    fn sync_safekeepers(&self, storage_auth_token: Option<String>) -> Result<Lsn> {
+    #[instrument(skip(self))]
+    fn sync_safekeepers(&self) -> Result<String> {
        let start_time = Utc::now();

        let sync_handle = Command::new(&self.pgbin)
            .args(["--sync-safekeepers"])
            .env("PGDATA", &self.pgdata) // we cannot use -D in this mode
-            .envs(if let Some(storage_auth_token) = &storage_auth_token {
+            .envs(if let Some(storage_auth_token) = &self.storage_auth_token {
                vec![("NEON_AUTH_TOKEN", storage_auth_token)]
            } else {
                vec![]
@@ -234,42 +201,45 @@ impl ComputeNode {
            );
        }

-        self.state.lock().unwrap().metrics.sync_safekeepers_ms = Utc::now()
-            .signed_duration_since(start_time)
-            .to_std()
-            .unwrap()
-            .as_millis() as u64;
+        self.metrics.sync_safekeepers_ms.store(
+            Utc::now()
+                .signed_duration_since(start_time)
+                .to_std()
+                .unwrap()
+                .as_millis() as u64,
+            Ordering::Relaxed,
+        );

-        let lsn = Lsn::from_str(String::from_utf8(sync_output.stdout)?.trim())?;
+        let lsn = String::from(String::from_utf8(sync_output.stdout)?.trim());

        Ok(lsn)
    }

    /// Do all the preparations like PGDATA directory creation, configuration,
    /// safekeepers sync, basebackup, etc.
-    #[instrument(skip(self, compute_state))]
-    pub fn prepare_pgdata(&self, compute_state: &ComputeState) -> Result<()> {
-        let pspec = compute_state.pspec.as_ref().expect("spec must be set");
+    #[instrument(skip(self))]
+    pub fn prepare_pgdata(&self) -> Result<()> {
+        let spec = &self.spec;
        let pgdata_path = Path::new(&self.pgdata);

        // Remove/create an empty pgdata directory and put configuration there.
        self.create_pgdata()?;
-        config::write_postgres_conf(&pgdata_path.join("postgresql.conf"), &pspec.spec)?;
+        config::write_postgres_conf(&pgdata_path.join("postgresql.conf"), spec)?;

        info!("starting safekeepers syncing");
        let lsn = self
-            .sync_safekeepers(pspec.storage_auth_token.clone())
+            .sync_safekeepers()
            .with_context(|| "failed to sync safekeepers")?;
        info!("safekeepers synced at LSN {}", lsn);

        info!(
            "getting basebackup@{} from pageserver {}",
-            lsn, &pspec.pageserver_connstr
+            lsn, &self.pageserver_connstr
        );
-        self.get_basebackup(compute_state, lsn).with_context(|| {
+        self.get_basebackup(&lsn).with_context(|| {
            format!(
                "failed to get basebackup@{} from pageserver {}",
-                lsn, &pspec.pageserver_connstr
+                lsn, &self.pageserver_connstr
            )
        })?;

@@ -282,16 +252,13 @@ impl ComputeNode {
    /// Start Postgres as a child process and manage DBs/roles.
    /// After that this will hang waiting on the postmaster process to exit.
    #[instrument(skip(self))]
-    pub fn start_postgres(
-        &self,
-        storage_auth_token: Option<String>,
-    ) -> Result<std::process::Child> {
+    pub fn start_postgres(&self) -> Result<std::process::Child> {
        let pgdata_path = Path::new(&self.pgdata);

        // Run postgres as a child process.
        let mut pg = Command::new(&self.pgbin)
            .args(["-D", &self.pgdata])
-            .envs(if let Some(storage_auth_token) = &storage_auth_token {
+            .envs(if let Some(storage_auth_token) = &self.storage_auth_token {
                vec![("NEON_AUTH_TOKEN", storage_auth_token)]
            } else {
                vec![]
@@ -304,9 +271,8 @@ impl ComputeNode {
        Ok(pg)
    }

-    /// Do initial configuration of the already started Postgres.
-    #[instrument(skip(self, compute_state))]
-    pub fn apply_config(&self, compute_state: &ComputeState) -> Result<()> {
+    #[instrument(skip(self))]
+    pub fn apply_config(&self) -> Result<()> {
        // If connection fails,
        // it may be the old node with `zenith_admin` superuser.
        //
@@ -337,62 +303,19 @@ impl ComputeNode {
        };

        // Proceed with post-startup configuration. Note, that order of operations is important.
-        let spec = &compute_state.pspec.as_ref().expect("spec must be set").spec;
-        handle_roles(spec, &mut client)?;
-        handle_databases(spec, &mut client)?;
-        handle_role_deletions(spec, self.connstr.as_str(), &mut client)?;
-        handle_grants(spec, self.connstr.as_str(), &mut client)?;
+        handle_roles(&self.spec, &mut client)?;
+        handle_databases(&self.spec, &mut client)?;
+        handle_role_deletions(self, &mut client)?;
+        handle_grants(self, &mut client)?;
        create_writability_check_data(&mut client)?;
-        handle_extensions(spec, &mut client)?;
+        handle_extensions(&self.spec, &mut client)?;

        // 'Close' connection
        drop(client);

        info!(
            "finished configuration of compute for project {}",
-            spec.cluster.cluster_id
-        );
-
-        Ok(())
-    }
-
-    // We could've wrapped this around `pg_ctl reload`, but right now we don't use
-    // `pg_ctl` for start / stop, so this just seems much easier to do as we already
-    // have opened connection to Postgres and superuser access.
-    #[instrument(skip(self, client))]
-    fn pg_reload_conf(&self, client: &mut Client) -> Result<()> {
-        client.simple_query("SELECT pg_reload_conf()")?;
-        Ok(())
-    }
-
-    /// Similar to `apply_config()`, but does a bit different sequence of operations,
-    /// as it's used to reconfigure a previously started and configured Postgres node.
-    #[instrument(skip(self))]
-    pub fn reconfigure(&self) -> Result<()> {
-        let spec = self.state.lock().unwrap().pspec.clone().unwrap().spec;
-
-        // Write new config
-        let pgdata_path = Path::new(&self.pgdata);
-        config::write_postgres_conf(&pgdata_path.join("postgresql.conf"), &spec)?;
-
-        let mut client = Client::connect(self.connstr.as_str(), NoTls)?;
-        self.pg_reload_conf(&mut client)?;
-
-        // Proceed with post-startup configuration. Note, that order of operations is important.
-        handle_roles(&spec, &mut client)?;
-        handle_databases(&spec, &mut client)?;
-        handle_role_deletions(&spec, self.connstr.as_str(), &mut client)?;
-        handle_grants(&spec, self.connstr.as_str(), &mut client)?;
-        handle_extensions(&spec, &mut client)?;
-
-        // 'Close' connection
-        drop(client);
-
-        let unknown_op = "unknown".to_string();
-        let op_id = spec.operation_uuid.as_ref().unwrap_or(&unknown_op);
-        info!(
-            "finished reconfiguration of compute node for operation {}",
-            op_id
+            self.spec.cluster.cluster_id
        );

        Ok(())
@@ -400,38 +323,40 @@ impl ComputeNode {

    #[instrument(skip(self))]
    pub fn start_compute(&self) -> Result<std::process::Child> {
-        let compute_state = self.state.lock().unwrap().clone();
-        let spec = compute_state.pspec.as_ref().expect("spec must be set");
        info!(
            "starting compute for project {}, operation {}, tenant {}, timeline {}",
-            spec.spec.cluster.cluster_id,
-            spec.spec.operation_uuid.as_deref().unwrap_or("None"),
-            spec.tenant_id,
-            spec.timeline_id,
+            self.spec.cluster.cluster_id,
+            self.spec.operation_uuid.as_ref().unwrap(),
+            self.tenant,
+            self.timeline,
        );

-        self.prepare_pgdata(&compute_state)?;
+        self.prepare_pgdata()?;

        let start_time = Utc::now();

-        let pg = self.start_postgres(spec.storage_auth_token.clone())?;
+        let pg = self.start_postgres()?;

-        self.apply_config(&compute_state)?;
+        self.apply_config()?;

        let startup_end_time = Utc::now();
-        {
-            let mut state = self.state.lock().unwrap();
-            state.metrics.config_ms = startup_end_time
+        self.metrics.config_ms.store(
+            startup_end_time
                .signed_duration_since(start_time)
                .to_std()
                .unwrap()
-                .as_millis() as u64;
-            state.metrics.total_startup_ms = startup_end_time
+                .as_millis() as u64,
+            Ordering::Relaxed,
+        );
+        self.metrics.total_startup_ms.store(
+            startup_end_time
                .signed_duration_since(self.start_time)
                .to_std()
                .unwrap()
-                .as_millis() as u64;
-        }
+                .as_millis() as u64,
+            Ordering::Relaxed,
+        );
+
        self.set_status(ComputeStatus::Running);

        Ok(pg)
--- a/compute_tools/src/config.rs
+++ b/compute_tools/src/config.rs
@@ -6,7 +6,7 @@ use std::path::Path;
 use anyhow::Result;

 use crate::pg_helpers::PgOptionsSerialize;
-use compute_api::spec::ComputeSpec;
+use crate::spec::ComputeSpec;

 /// Check that `line` is inside a text file and put it there if it is not.
 /// Create file if it doesn't exist.
--- a/compute_tools/src/configurator.rs
+++ b/compute_tools/src/configurator.rs
@@ -1,54 +0,0 @@
-use std::sync::Arc;
-use std::thread;
-
-use anyhow::Result;
-use tracing::{error, info, instrument};
-
-use compute_api::responses::ComputeStatus;
-
-use crate::compute::ComputeNode;
-
-#[instrument(skip(compute))]
-fn configurator_main_loop(compute: &Arc<ComputeNode>) {
-    info!("waiting for reconfiguration requests");
-    loop {
-        let state = compute.state.lock().unwrap();
-        let mut state = compute.state_changed.wait(state).unwrap();
-
-        if state.status == ComputeStatus::ConfigurationPending {
-            info!("got configuration request");
-            state.status = ComputeStatus::Configuration;
-            compute.state_changed.notify_all();
-            drop(state);
-
-            let mut new_status = ComputeStatus::Failed;
-            if let Err(e) = compute.reconfigure() {
-                error!("could not configure compute node: {}", e);
-            } else {
-                new_status = ComputeStatus::Running;
-                info!("compute node configured");
-            }
-
-            // XXX: used to test that API is blocking
-            // std::thread::sleep(std::time::Duration::from_millis(10000));
-
-            compute.set_status(new_status);
-        } else if state.status == ComputeStatus::Failed {
-            info!("compute node is now in Failed state, exiting");
-            break;
-        } else {
-            info!("woken up for compute status: {:?}, sleeping", state.status);
-        }
-    }
-}
-
-pub fn launch_configurator(compute: &Arc<ComputeNode>) -> Result<thread::JoinHandle<()>> {
-    let compute = Arc::clone(compute);
-
-    Ok(thread::Builder::new()
-        .name("compute-configurator".into())
-        .spawn(move || {
-            configurator_main_loop(&compute);
-            info!("configurator thread is exited");
-        })?)
-}
--- a/compute_tools/src/http/api.rs
+++ b/compute_tools/src/http/api.rs
@@ -3,35 +3,15 @@ use std::net::SocketAddr;
 use std::sync::Arc;
 use std::thread;

-use crate::compute::{ComputeNode, ComputeState, ParsedSpec};
-use compute_api::requests::ConfigurationRequest;
-use compute_api::responses::{ComputeStatus, ComputeStatusResponse, GenericAPIError};
-
+use crate::compute::ComputeNode;
 use anyhow::Result;
 use hyper::service::{make_service_fn, service_fn};
 use hyper::{Body, Method, Request, Response, Server, StatusCode};
 use num_cpus;
 use serde_json;
-use tokio::task;
 use tracing::{error, info};
 use tracing_utils::http::OtelName;

-fn status_response_from_state(state: &ComputeState) -> ComputeStatusResponse {
-    ComputeStatusResponse {
-        tenant: state
-            .pspec
-            .as_ref()
-            .map(|pspec| pspec.tenant_id.to_string()),
-        timeline: state
-            .pspec
-            .as_ref()
-            .map(|pspec| pspec.timeline_id.to_string()),
-        status: state.status,
-        last_active: state.last_active,
-        error: state.error.clone(),
-    }
-}
-
 // Service function to handle all available routes.
 async fn routes(req: Request<Body>, compute: &Arc<ComputeNode>) -> Response<Body> {
    //
@@ -43,45 +23,26 @@ async fn routes(req: Request<Body>, compute: &Arc<ComputeNode>) -> Response<Body
        // Serialized compute state.
        (&Method::GET, "/status") => {
            info!("serving /status GET request");
-            let state = compute.state.lock().unwrap();
-            let status_response = status_response_from_state(&state);
-            Response::new(Body::from(serde_json::to_string(&status_response).unwrap()))
+            let state = compute.state.read().unwrap();
+            Response::new(Body::from(serde_json::to_string(&*state).unwrap()))
        }

        // Startup metrics in JSON format. Keep /metrics reserved for a possible
        // future use for Prometheus metrics format.
        (&Method::GET, "/metrics.json") => {
            info!("serving /metrics.json GET request");
-            let metrics = compute.state.lock().unwrap().metrics.clone();
-            Response::new(Body::from(serde_json::to_string(&metrics).unwrap()))
+            Response::new(Body::from(serde_json::to_string(&compute.metrics).unwrap()))
        }

        // Collect Postgres current usage insights
        (&Method::GET, "/insights") => {
            info!("serving /insights GET request");
-            let status = compute.get_status();
-            if status != ComputeStatus::Running {
-                let msg = format!("compute is not running, current status: {:?}", status);
-                error!(msg);
-                return Response::new(Body::from(msg));
-            }
-
            let insights = compute.collect_insights().await;
            Response::new(Body::from(insights))
        }

        (&Method::POST, "/check_writability") => {
            info!("serving /check_writability POST request");
-            let status = compute.get_status();
-            if status != ComputeStatus::Running {
-                let msg = format!(
-                    "invalid compute status for check_writability request: {:?}",
-                    status
-                );
-                error!(msg);
-                return Response::new(Body::from(msg));
-            }
-
            let res = crate::checker::check_writability(compute).await;
            match res {
                Ok(_) => Response::new(Body::from("true")),
@@ -100,23 +61,6 @@ async fn routes(req: Request<Body>, compute: &Arc<ComputeNode>) -> Response<Body
            ))
        }

-        // Accept spec in JSON format and request compute configuration. If
-        // anything goes wrong after we set the compute status to `ConfigurationPending`
-        // and update compute state with new spec, we basically leave compute
-        // in the potentially wrong state. That said, it's control-plane's
-        // responsibility to watch compute state after reconfiguration request
-        // and to clean restart in case of errors.
-        (&Method::POST, "/configure") => {
-            info!("serving /configure POST request");
-            match handle_configure_request(req, compute).await {
-                Ok(msg) => Response::new(Body::from(msg)),
-                Err((msg, code)) => {
-                    error!("error handling /configure request: {msg}");
-                    render_json_error(&msg, code)
-                }
-            }
-        }
-
        // Return the `404 Not Found` for any other routes.
        _ => {
            let mut not_found = Response::new(Body::from("404 Not Found"));
@@ -126,94 +70,6 @@ async fn routes(req: Request<Body>, compute: &Arc<ComputeNode>) -> Response<Body
    }
 }

-async fn handle_configure_request(
-    req: Request<Body>,
-    compute: &Arc<ComputeNode>,
-) -> Result<String, (String, StatusCode)> {
-    if !compute.live_config_allowed {
-        return Err((
-            "live configuration is not allowed for this compute node".to_string(),
-            StatusCode::PRECONDITION_FAILED,
-        ));
-    }
-
-    let body_bytes = hyper::body::to_bytes(req.into_body()).await.unwrap();
-    let spec_raw = String::from_utf8(body_bytes.to_vec()).unwrap();
-    if let Ok(request) = serde_json::from_str::<ConfigurationRequest>(&spec_raw) {
-        let spec = request.spec;
-
-        let parsed_spec = match ParsedSpec::try_from(spec) {
-            Ok(ps) => ps,
-            Err(msg) => return Err((msg, StatusCode::PRECONDITION_FAILED)),
-        };
-
-        // XXX: wrap state update under lock in code blocks. Otherwise,
-        // we will try to `Send` `mut state` into the spawned thread
-        // bellow, which will cause error:
-        // ```
-        // error: future cannot be sent between threads safely
-        // ```
-        {
-            let mut state = compute.state.lock().unwrap();
-            if state.status != ComputeStatus::Empty && state.status != ComputeStatus::Running {
-                let msg = format!(
-                    "invalid compute status for configuration request: {:?}",
-                    state.status.clone()
-                );
-                return Err((msg, StatusCode::PRECONDITION_FAILED));
-            }
-            state.pspec = Some(parsed_spec);
-            state.status = ComputeStatus::ConfigurationPending;
-            compute.state_changed.notify_all();
-            drop(state);
-            info!("set new spec and notified waiters");
-        }
-
-        // Spawn a blocking thread to wait for compute to become Running.
-        // This is needed to do not block the main pool of workers and
-        // be able to serve other requests while some particular request
-        // is waiting for compute to finish configuration.
-        let c = compute.clone();
-        task::spawn_blocking(move || {
-            let mut state = c.state.lock().unwrap();
-            while state.status != ComputeStatus::Running {
-                state = c.state_changed.wait(state).unwrap();
-                info!(
-                    "waiting for compute to become Running, current status: {:?}",
-                    state.status
-                );
-
-                if state.status == ComputeStatus::Failed {
-                    let err = state.error.as_ref().map_or("unknown error", |x| x);
-                    let msg = format!("compute configuration failed: {:?}", err);
-                    return Err((msg, StatusCode::INTERNAL_SERVER_ERROR));
-                }
-            }
-
-            Ok(())
-        })
-        .await
-        .unwrap()?;
-
-        // Return current compute state if everything went well.
-        let state = compute.state.lock().unwrap().clone();
-        let status_response = status_response_from_state(&state);
-        Ok(serde_json::to_string(&status_response).unwrap())
-    } else {
-        Err(("invalid spec".to_string(), StatusCode::BAD_REQUEST))
-    }
-}
-
-fn render_json_error(e: &str, status: StatusCode) -> Response<Body> {
-    let error = GenericAPIError {
-        error: e.to_string(),
-    };
-    Response::builder()
-        .status(status)
-        .body(Body::from(serde_json::to_string(&error).unwrap()))
-        .unwrap()
-}
-
 // Main Hyper HTTP server function that runs it and blocks waiting on it forever.
 #[tokio::main]
 async fn serve(state: Arc<ComputeNode>) {
--- a/compute_tools/src/http/openapi_spec.yaml
+++ b/compute_tools/src/http/openapi_spec.yaml
@@ -11,7 +11,7 @@ paths:
    get:
      tags:
      - Info
-      summary: Get compute node internal status.
+      summary: Get compute node internal status
      description: ""
      operationId: getComputeStatus
      responses:
@@ -26,7 +26,7 @@ paths:
    get:
      tags:
      - Info
-      summary: Get compute node startup metrics in JSON format.
+      summary: Get compute node startup metrics in JSON format
      description: ""
      operationId: getComputeMetricsJSON
      responses:
@@ -41,9 +41,9 @@ paths:
    get:
      tags:
      - Info
-      summary: Get current compute insights in JSON format.
+      summary: Get current compute insights in JSON format
      description: |
-        Note, that this doesn't include any historical data.
+        Note, that this doesn't include any historical data
      operationId: getComputeInsights
      responses:
        200:
@@ -56,12 +56,12 @@ paths:
  /info:
    get:
      tags:
-      - Info
-      summary: Get info about the compute pod / VM.
+      - "info"
+      summary: Get info about the compute Pod/VM
      description: ""
      operationId: getInfo
      responses:
-        200:
+        "200":
          description: Info
          content:
            application/json:
@@ -72,7 +72,7 @@ paths:
    post:
      tags:
      - Check
-      summary: Check that we can write new data on this compute.
+      summary: Check that we can write new data on this compute
      description: ""
      operationId: checkComputeWritability
      responses:
@@ -82,64 +82,9 @@ paths:
            text/plain:
              schema:
                type: string
-                description: Error text or 'true' if check passed.
+                description: Error text or 'true' if check passed
                example: "true"

-  /configure:
-    post:
-      tags:
-      - Configure
-      summary: Perform compute node configuration.
-      description: |
-        This is a blocking API endpoint, i.e. it blocks waiting until
-        compute is finished configuration and is in `Running` state.
-        Optional non-blocking mode could be added later.
-      operationId: configureCompute
-      requestBody:
-        description: Configuration request.
-        required: true
-        content:
-          application/json:
-            schema:
-              type: object
-              required:
-                - spec
-              properties:
-                spec:
-                  # XXX: I don't want to explain current spec in the OpenAPI format,
-                  # as it could be changed really soon. Consider doing it later.
-                  type: object
-      responses:
-        200:
-          description: Compute configuration finished.
-          content:
-            application/json:
-              schema:
-                $ref: "#/components/schemas/ComputeState"
-        400:
-          description: Provided spec is invalid.
-          content:
-            application/json:
-              schema:
-                $ref: "#/components/schemas/GenericError"
-        412:
-          description: |
-            It's not possible to do live-configuration of the compute.
-            It's either in the wrong state, or compute doesn't use pull
-            mode of configuration.
-          content:
-            application/json:
-              schema:
-                $ref: "#/components/schemas/GenericError"
-        500:
-          description: |
-            Compute configuration request was processed, but error
-            occurred. Compute will likely shutdown soon.
-          content:
-            application/json:
-              schema:
-                $ref: "#/components/schemas/GenericError"
-
 components:
  securitySchemes:
    JWT:
@@ -150,7 +95,7 @@ components:
  schemas:
    ComputeMetrics:
      type: object
-      description: Compute startup metrics.
+      description: Compute startup metrics
      required:
        - sync_safekeepers_ms
        - basebackup_ms
@@ -168,7 +113,7 @@ components:

    Info:
      type: object
-      description: Information about VM/Pod.
+      description: Information about VM/Pod
      required:
        - num_cpus
      properties:
@@ -185,26 +130,17 @@ components:
          $ref: '#/components/schemas/ComputeStatus'
        last_active:
          type: string
-          description: The last detected compute activity timestamp in UTC and RFC3339 format.
+          description: The last detected compute activity timestamp in UTC and RFC3339 format
          example: "2022-10-12T07:20:50.52Z"
        error:
          type: string
-          description: Text of the error during compute startup, if any.
-          example: ""
-        tenant:
-          type: string
-          description: Identifier of the current tenant served by compute node, if any.
-          example: c9269c359e9a199fad1ea0981246a78f
-        timeline:
-          type: string
-          description: Identifier of the current timeline served by compute node, if any.
-          example: ece7de74d4b8cbe5433a68ce4d1b97b4
+          description: Text of the error during compute startup, if any

    ComputeInsights:
      type: object
      properties:
        pg_stat_statements:
-          description: Contains raw output from pg_stat_statements in JSON format.
+          description: Contains raw output from pg_stat_statements in JSON format
          type: array
          items:
            type: object
@@ -215,19 +151,6 @@ components:
        - init
        - failed
        - running
-      example: running
-
-    #
-    # Errors
-    #
-
-    GenericError:
-      type: object
-      required:
-        - error
-      properties:
-        error:
-          type: string

 security:
  - JWT: []
--- a/compute_tools/src/lib.rs
+++ b/compute_tools/src/lib.rs
@@ -4,7 +4,6 @@
 //!
 pub mod checker;
 pub mod config;
-pub mod configurator;
 pub mod http;
 #[macro_use]
 pub mod logger;
--- a/compute_tools/src/monitor.rs
+++ b/compute_tools/src/monitor.rs
@@ -46,7 +46,7 @@ fn watch_compute_activity(compute: &ComputeNode) {
                            AND usename != 'cloud_admin';", // XXX: find a better way to filter other monitors?
                        &[],
                    );
-                let mut last_active = compute.state.lock().unwrap().last_active;
+                let mut last_active = compute.state.read().unwrap().last_active;

                if let Ok(backs) = backends {
                    let mut idle_backs: Vec<DateTime<Utc>> = vec![];
@@ -87,7 +87,7 @@ fn watch_compute_activity(compute: &ComputeNode) {
                }

                // Update the last activity in the shared state if we got a more recent one.
-                let mut state = compute.state.lock().unwrap();
+                let mut state = compute.state.write().unwrap();
                if last_active > state.last_active {
                    state.last_active = last_active;
                    debug!("set the last compute activity time to: {}", last_active);
--- a/compute_tools/src/pg_helpers.rs
+++ b/compute_tools/src/pg_helpers.rs
@@ -10,12 +10,43 @@ use std::time::{Duration, Instant};
 use anyhow::{bail, Result};
 use notify::{RecursiveMode, Watcher};
 use postgres::{Client, Transaction};
+use serde::Deserialize;
 use tracing::{debug, instrument};

-use compute_api::spec::{Database, GenericOption, GenericOptions, PgIdent, Role};
-
 const POSTGRES_WAIT_TIMEOUT: Duration = Duration::from_millis(60 * 1000); // milliseconds

+/// Rust representation of Postgres role info with only those fields
+/// that matter for us.
+#[derive(Clone, Deserialize)]
+pub struct Role {
+    pub name: PgIdent,
+    pub encrypted_password: Option<String>,
+    pub options: GenericOptions,
+}
+
+/// Rust representation of Postgres database info with only those fields
+/// that matter for us.
+#[derive(Clone, Deserialize)]
+pub struct Database {
+    pub name: PgIdent,
+    pub owner: PgIdent,
+    pub options: GenericOptions,
+}
+
+/// Common type representing both SQL statement params with or without value,
+/// like `LOGIN` or `OWNER username` in the `CREATE/ALTER ROLE`, and config
+/// options like `wal_level = logical`.
+#[derive(Clone, Deserialize)]
+pub struct GenericOption {
+    pub name: String,
+    pub value: Option<String>,
+    pub vartype: String,
+}
+
+/// Optional collection of `GenericOption`'s. Type alias allows us to
+/// declare a `trait` on it.
+pub type GenericOptions = Option<Vec<GenericOption>>;
+
 /// Escape a string for including it in a SQL literal
 fn escape_literal(s: &str) -> String {
    s.replace('\'', "''").replace('\\', "\\\\")
@@ -27,14 +58,9 @@ fn escape_conf_value(s: &str) -> String {
    s.replace('\'', "''").replace('\\', "\\\\")
 }

-trait GenericOptionExt {
-    fn to_pg_option(&self) -> String;
-    fn to_pg_setting(&self) -> String;
-}
-
-impl GenericOptionExt for GenericOption {
+impl GenericOption {
    /// Represent `GenericOption` as SQL statement parameter.
-    fn to_pg_option(&self) -> String {
+    pub fn to_pg_option(&self) -> String {
        if let Some(val) = &self.value {
            match self.vartype.as_ref() {
                "string" => format!("{} '{}'", self.name, escape_literal(val)),
@@ -46,7 +72,7 @@ impl GenericOptionExt for GenericOption {
    }

    /// Represent `GenericOption` as configuration option.
-    fn to_pg_setting(&self) -> String {
+    pub fn to_pg_setting(&self) -> String {
        if let Some(val) = &self.value {
            match self.vartype.as_ref() {
                "string" => format!("{} = '{}'", self.name, escape_conf_value(val)),
@@ -105,14 +131,10 @@ impl GenericOptionsSearch for GenericOptions {
    }
 }

-pub trait RoleExt {
-    fn to_pg_options(&self) -> String;
-}
-
-impl RoleExt for Role {
+impl Role {
    /// Serialize a list of role parameters into a Postgres-acceptable
    /// string of arguments.
-    fn to_pg_options(&self) -> String {
+    pub fn to_pg_options(&self) -> String {
        // XXX: consider putting LOGIN as a default option somewhere higher, e.g. in control-plane.
        // For now, we do not use generic `options` for roles. Once used, add
        // `self.options.as_pg_options()` somewhere here.
@@ -137,17 +159,21 @@ impl RoleExt for Role {
    }
 }

-pub trait DatabaseExt {
-    fn to_pg_options(&self) -> String;
-}
+impl Database {
+    pub fn new(name: PgIdent, owner: PgIdent) -> Self {
+        Self {
+            name,
+            owner,
+            options: None,
+        }
+    }

-impl DatabaseExt for Database {
    /// Serialize a list of database parameters into a Postgres-acceptable
    /// string of arguments.
    /// NB: `TEMPLATE` is actually also an identifier, but so far we only need
    /// to use `template0` and `template1`, so it is not a problem. Yet in the future
    /// it may require a proper quoting too.
-    fn to_pg_options(&self) -> String {
+    pub fn to_pg_options(&self) -> String {
        let mut params: String = self.options.as_pg_options();
        write!(params, " OWNER {}", &self.owner.pg_quote())
            .expect("String is documented to not to error during write operations");
@@ -156,6 +182,10 @@ impl DatabaseExt for Database {
    }
 }

+/// String type alias representing Postgres identifier and
+/// intended to be used for DB / role names.
+pub type PgIdent = String;
+
 /// Generic trait used to provide quoting / encoding for strings used in the
 /// Postgres SQL queries and DATABASE_URL.
 pub trait Escaping {
@@ -196,11 +226,7 @@ pub fn get_existing_dbs(client: &mut Client) -> Result<Vec<Database>> {
            &[],
        )?
        .iter()
-        .map(|row| Database {
-            name: row.get("datname"),
-            owner: row.get("owner"),
-            options: None,
-        })
+        .map(|row| Database::new(row.get("datname"), row.get("owner")))
        .collect();

    Ok(postgres_dbs)
--- a/compute_tools/src/spec.rs
+++ b/compute_tools/src/spec.rs
@@ -1,45 +1,57 @@
+use std::collections::HashMap;
 use std::path::Path;
 use std::str::FromStr;

-use anyhow::{anyhow, bail, Result};
+use anyhow::Result;
 use postgres::config::Config;
 use postgres::{Client, NoTls};
+use serde::Deserialize;
 use tracing::{info, info_span, instrument, span_enabled, warn, Level};

+use crate::compute::ComputeNode;
 use crate::config;
 use crate::params::PG_HBA_ALL_MD5;
 use crate::pg_helpers::*;

-use compute_api::responses::ControlPlaneSpecResponse;
-use compute_api::spec::{ComputeSpec, Database, PgIdent, Role};
+/// Cluster spec or configuration represented as an optional number of
+/// delta operations + final cluster state description.
+#[derive(Clone, Deserialize)]
+pub struct ComputeSpec {
+    pub format_version: f32,
+    pub timestamp: String,
+    pub operation_uuid: Option<String>,
+    /// Expected cluster state at the end of transition process.
+    pub cluster: Cluster,
+    pub delta_operations: Option<Vec<DeltaOp>>,

-/// Request spec from the control-plane by compute_id. If `NEON_CONSOLE_JWT`
-/// env variable is set, it will be used for authorization.
-pub fn get_spec_from_control_plane(base_uri: &str, compute_id: &str) -> Result<ComputeSpec> {
-    let cp_uri = format!("{base_uri}/management/api/v2/computes/{compute_id}/spec");
-    let jwt: String = match std::env::var("NEON_CONSOLE_JWT") {
-        Ok(v) => v,
-        Err(_) => "".to_string(),
-    };
-    info!("getting spec from control plane: {}", cp_uri);
+    pub storage_auth_token: Option<String>,

-    // TODO: check the response. We should distinguish cases when it's
-    // - network error, then retry
-    // - no spec for compute yet, then wait
-    // - compute id is unknown or any other error, then bail out
-    let resp: ControlPlaneSpecResponse = reqwest::blocking::Client::new()
-        .get(cp_uri)
-        .header("Authorization", jwt)
-        .send()
-        .map_err(|e| anyhow!("could not send spec request to control plane: {}", e))?
-        .json()
-        .map_err(|e| anyhow!("could not get compute spec from control plane: {}", e))?;
+    pub startup_tracing_context: Option<HashMap<String, String>>,
+}

-    if let Some(spec) = resp.spec {
-        Ok(spec)
-    } else {
-        bail!("could not get compute spec from control plane")
-    }
+/// Cluster state seen from the perspective of the external tools
+/// like Rails web console.
+#[derive(Clone, Deserialize)]
+pub struct Cluster {
+    pub cluster_id: String,
+    pub name: String,
+    pub state: Option<String>,
+    pub roles: Vec<Role>,
+    pub databases: Vec<Database>,
+    pub settings: GenericOptions,
+}
+
+/// Single cluster state changing operation that could not be represented as
+/// a static `Cluster` structure. For example:
+/// - DROP DATABASE
+/// - DROP ROLE
+/// - ALTER ROLE name RENAME TO new_name
+/// - ALTER DATABASE name RENAME TO new_name
+#[derive(Clone, Deserialize)]
+pub struct DeltaOp {
+    pub action: String,
+    pub name: PgIdent,
+    pub new_name: Option<PgIdent>,
 }

 /// It takes cluster specification and does the following:
@@ -214,8 +226,8 @@ pub fn handle_roles(spec: &ComputeSpec, client: &mut Client) -> Result<()> {

 /// Reassign all dependent objects and delete requested roles.
 #[instrument(skip_all)]
-pub fn handle_role_deletions(spec: &ComputeSpec, connstr: &str, client: &mut Client) -> Result<()> {
-    if let Some(ops) = &spec.delta_operations {
+pub fn handle_role_deletions(node: &ComputeNode, client: &mut Client) -> Result<()> {
+    if let Some(ops) = &node.spec.delta_operations {
        // First, reassign all dependent objects to db owners.
        info!("reassigning dependent objects of to-be-deleted roles");

@@ -232,7 +244,7 @@ pub fn handle_role_deletions(spec: &ComputeSpec, connstr: &str, client: &mut Cli
            // Check that role is still present in Postgres, as this could be a
            // restart with the same spec after role deletion.
            if op.action == "delete_role" && existing_roles.iter().any(|r| r.name == op.name) {
-                reassign_owned_objects(spec, connstr, &op.name)?;
+                reassign_owned_objects(node, &op.name)?;
            }
        }

@@ -256,10 +268,10 @@ pub fn handle_role_deletions(spec: &ComputeSpec, connstr: &str, client: &mut Cli
 }

 // Reassign all owned objects in all databases to the owner of the database.
-fn reassign_owned_objects(spec: &ComputeSpec, connstr: &str, role_name: &PgIdent) -> Result<()> {
-    for db in &spec.cluster.databases {
+fn reassign_owned_objects(node: &ComputeNode, role_name: &PgIdent) -> Result<()> {
+    for db in &node.spec.cluster.databases {
        if db.owner != *role_name {
-            let mut conf = Config::from_str(connstr)?;
+            let mut conf = Config::from_str(node.connstr.as_str())?;
            conf.dbname(&db.name);

            let mut client = conf.connect(NoTls)?;
@@ -404,7 +416,9 @@ pub fn handle_databases(spec: &ComputeSpec, client: &mut Client) -> Result<()> {
 /// Grant CREATE ON DATABASE to the database owner and do some other alters and grants
 /// to allow users creating trusted extensions and re-creating `public` schema, for example.
 #[instrument(skip_all)]
-pub fn handle_grants(spec: &ComputeSpec, connstr: &str, client: &mut Client) -> Result<()> {
+pub fn handle_grants(node: &ComputeNode, client: &mut Client) -> Result<()> {
+    let spec = &node.spec;
+
    info!("cluster spec grants:");

    // We now have a separate `web_access` role to connect to the database
@@ -436,8 +450,8 @@ pub fn handle_grants(spec: &ComputeSpec, connstr: &str, client: &mut Client) ->
    // Do some per-database access adjustments. We'd better do this at db creation time,
    // but CREATE DATABASE isn't transactional. So we cannot create db + do some grants
    // atomically.
-    for db in &spec.cluster.databases {
-        let mut conf = Config::from_str(connstr)?;
+    for db in &node.spec.cluster.databases {
+        let mut conf = Config::from_str(node.connstr.as_str())?;
        conf.dbname(&db.name);

        let mut db_client = conf.connect(NoTls)?;
--- a/libs/compute_api/tests/cluster_spec.json
+++ b/libs/compute_api/tests/cluster_spec.json
--- a/compute_tools/tests/pg_helpers_tests.rs
+++ b/compute_tools/tests/pg_helpers_tests.rs
@@ -1,13 +1,14 @@
 #[cfg(test)]
 mod pg_helpers_tests {
+
    use std::fs::File;

-    use compute_api::spec::{ComputeSpec, GenericOption, GenericOptions, PgIdent};
    use compute_tools::pg_helpers::*;
+    use compute_tools::spec::ComputeSpec;

    #[test]
    fn params_serialize() {
-        let file = File::open("../libs/compute_api/tests/cluster_spec.json").unwrap();
+        let file = File::open("tests/cluster_spec.json").unwrap();
        let spec: ComputeSpec = serde_json::from_reader(file).unwrap();

        assert_eq!(
@@ -22,7 +23,7 @@ mod pg_helpers_tests {

    #[test]
    fn settings_serialize() {
-        let file = File::open("../libs/compute_api/tests/cluster_spec.json").unwrap();
+        let file = File::open("tests/cluster_spec.json").unwrap();
        let spec: ComputeSpec = serde_json::from_reader(file).unwrap();

        assert_eq!(
--- a/control_plane/src/bin/neon_local.rs
+++ b/control_plane/src/bin/neon_local.rs
@@ -7,7 +7,7 @@
 //!
 use anyhow::{anyhow, bail, Context, Result};
 use clap::{value_parser, Arg, ArgAction, ArgMatches, Command};
-use control_plane::endpoint::ComputeControlPlane;
+use control_plane::compute::ComputeControlPlane;
 use control_plane::local_env::LocalEnv;
 use control_plane::pageserver::PageServerNode;
 use control_plane::safekeeper::SafekeeperNode;
@@ -106,9 +106,8 @@ fn main() -> Result<()> {
            "start" => handle_start_all(sub_args, &env),
            "stop" => handle_stop_all(sub_args, &env),
            "pageserver" => handle_pageserver(sub_args, &env),
+            "pg" => handle_pg(sub_args, &env),
            "safekeeper" => handle_safekeeper(sub_args, &env),
-            "endpoint" => handle_endpoint(sub_args, &env),
-            "pg" => bail!("'pg' subcommand has been renamed to 'endpoint'"),
            _ => bail!("unexpected subcommand {sub_name}"),
        };

@@ -471,10 +470,10 @@ fn handle_timeline(timeline_match: &ArgMatches, env: &mut local_env::LocalEnv) -
            let mut cplane = ComputeControlPlane::load(env.clone())?;
            println!("Importing timeline into pageserver ...");
            pageserver.timeline_import(tenant_id, timeline_id, base, pg_wal, pg_version)?;
+            println!("Creating node for imported timeline ...");
            env.register_branch_mapping(name.to_string(), tenant_id, timeline_id)?;

-            println!("Creating endpoint for imported timeline ...");
-            cplane.new_endpoint(tenant_id, name, timeline_id, None, None, pg_version)?;
+            cplane.new_node(tenant_id, name, timeline_id, None, None, pg_version)?;
            println!("Done");
        }
        Some(("branch", branch_match)) => {
@@ -522,10 +521,10 @@ fn handle_timeline(timeline_match: &ArgMatches, env: &mut local_env::LocalEnv) -
    Ok(())
 }

-fn handle_endpoint(ep_match: &ArgMatches, env: &local_env::LocalEnv) -> Result<()> {
-    let (sub_name, sub_args) = match ep_match.subcommand() {
-        Some(ep_subcommand_data) => ep_subcommand_data,
-        None => bail!("no endpoint subcommand provided"),
+fn handle_pg(pg_match: &ArgMatches, env: &local_env::LocalEnv) -> Result<()> {
+    let (sub_name, sub_args) = match pg_match.subcommand() {
+        Some(pg_subcommand_data) => pg_subcommand_data,
+        None => bail!("no pg subcommand provided"),
    };

    let mut cplane = ComputeControlPlane::load(env.clone())?;
@@ -547,7 +546,7 @@ fn handle_endpoint(ep_match: &ArgMatches, env: &local_env::LocalEnv) -> Result<(
            table.load_preset(comfy_table::presets::NOTHING);

            table.set_header([
-                "ENDPOINT",
+                "NODE",
                "ADDRESS",
                "TIMELINE",
                "BRANCH NAME",
@@ -555,39 +554,39 @@ fn handle_endpoint(ep_match: &ArgMatches, env: &local_env::LocalEnv) -> Result<(
                "STATUS",
            ]);

-            for (endpoint_id, endpoint) in cplane
-                .endpoints
+            for ((_, node_name), node) in cplane
+                .nodes
                .iter()
-                .filter(|(_, endpoint)| endpoint.tenant_id == tenant_id)
+                .filter(|((node_tenant_id, _), _)| node_tenant_id == &tenant_id)
            {
-                let lsn_str = match endpoint.lsn {
+                let lsn_str = match node.lsn {
                    None => {
-                        // -> primary endpoint
+                        // -> primary node
                        // Use the LSN at the end of the timeline.
                        timeline_infos
-                            .get(&endpoint.timeline_id)
+                            .get(&node.timeline_id)
                            .map(|bi| bi.last_record_lsn.to_string())
                            .unwrap_or_else(|| "?".to_string())
                    }
                    Some(lsn) => {
-                        // -> read-only endpoint
-                        // Use the endpoint's LSN.
+                        // -> read-only node
+                        // Use the node's LSN.
                        lsn.to_string()
                    }
                };

                let branch_name = timeline_name_mappings
-                    .get(&TenantTimelineId::new(tenant_id, endpoint.timeline_id))
+                    .get(&TenantTimelineId::new(tenant_id, node.timeline_id))
                    .map(|name| name.as_str())
                    .unwrap_or("?");

                table.add_row([
-                    endpoint_id.as_str(),
-                    &endpoint.address.to_string(),
-                    &endpoint.timeline_id.to_string(),
+                    node_name.as_str(),
+                    &node.address.to_string(),
+                    &node.timeline_id.to_string(),
                    branch_name,
                    lsn_str.as_str(),
-                    endpoint.status(),
+                    node.status(),
                ]);
            }

@@ -598,10 +597,10 @@ fn handle_endpoint(ep_match: &ArgMatches, env: &local_env::LocalEnv) -> Result<(
                .get_one::<String>("branch-name")
                .map(|s| s.as_str())
                .unwrap_or(DEFAULT_BRANCH_NAME);
-            let endpoint_id = sub_args
-                .get_one::<String>("endpoint_id")
-                .map(String::to_string)
-                .unwrap_or_else(|| format!("ep-{branch_name}"));
+            let node_name = sub_args
+                .get_one::<String>("node")
+                .map(|node_name| node_name.to_string())
+                .unwrap_or_else(|| format!("{branch_name}_node"));

            let lsn = sub_args
                .get_one::<String>("lsn")
@@ -619,15 +618,17 @@ fn handle_endpoint(ep_match: &ArgMatches, env: &local_env::LocalEnv) -> Result<(
                .copied()
                .context("Failed to parse postgres version from the argument string")?;

-            cplane.new_endpoint(tenant_id, &endpoint_id, timeline_id, lsn, port, pg_version)?;
+            let node =
+                cplane.new_node(tenant_id, &node_name, timeline_id, lsn, port, pg_version)?;
+            println!("{}", node.pgdata().display());
        }
        "start" => {
            let port: Option<u16> = sub_args.get_one::<u16>("port").copied();
-            let endpoint_id = sub_args
-                .get_one::<String>("endpoint_id")
-                .ok_or_else(|| anyhow!("No endpoint ID was provided to start"))?;
+            let node_name = sub_args
+                .get_one::<String>("node")
+                .ok_or_else(|| anyhow!("No node name was provided to start"))?;

-            let endpoint = cplane.endpoints.get(endpoint_id.as_str());
+            let node = cplane.nodes.get(&(tenant_id, node_name.to_string()));

            let auth_token = if matches!(env.pageserver.pg_auth_type, AuthType::NeonJWT) {
                let claims = Claims::new(Some(tenant_id), Scope::Tenant);
@@ -637,9 +638,9 @@ fn handle_endpoint(ep_match: &ArgMatches, env: &local_env::LocalEnv) -> Result<(
                None
            };

-            if let Some(endpoint) = endpoint {
-                println!("Starting existing endpoint {endpoint_id}...");
-                endpoint.start(&auth_token)?;
+            if let Some(node) = node {
+                println!("Starting existing postgres {node_name}...");
+                node.start(&auth_token)?;
            } else {
                let branch_name = sub_args
                    .get_one::<String>("branch-name")
@@ -664,33 +665,27 @@ fn handle_endpoint(ep_match: &ArgMatches, env: &local_env::LocalEnv) -> Result<(
                // start --port X
                // stop
                // start <-- will also use port X even without explicit port argument
-                println!("Starting new endpoint {endpoint_id} (PostgreSQL v{pg_version}) on timeline {timeline_id} ...");
+                println!("Starting new postgres (v{pg_version}) {node_name} on timeline {timeline_id} ...");

-                let ep = cplane.new_endpoint(
-                    tenant_id,
-                    endpoint_id,
-                    timeline_id,
-                    lsn,
-                    port,
-                    pg_version,
-                )?;
-                ep.start(&auth_token)?;
+                let node =
+                    cplane.new_node(tenant_id, node_name, timeline_id, lsn, port, pg_version)?;
+                node.start(&auth_token)?;
            }
        }
        "stop" => {
-            let endpoint_id = sub_args
-                .get_one::<String>("endpoint_id")
-                .ok_or_else(|| anyhow!("No endpoint ID was provided to stop"))?;
+            let node_name = sub_args
+                .get_one::<String>("node")
+                .ok_or_else(|| anyhow!("No node name was provided to stop"))?;
            let destroy = sub_args.get_flag("destroy");

-            let endpoint = cplane
-                .endpoints
-                .get(endpoint_id.as_str())
-                .with_context(|| format!("postgres endpoint {endpoint_id} is not found"))?;
-            endpoint.stop(destroy)?;
+            let node = cplane
+                .nodes
+                .get(&(tenant_id, node_name.to_string()))
+                .with_context(|| format!("postgres {node_name} is not found"))?;
+            node.stop(destroy)?;
        }

-        _ => bail!("Unexpected endpoint subcommand '{sub_name}'"),
+        _ => bail!("Unexpected pg subcommand '{sub_name}'"),
    }

    Ok(())
@@ -809,7 +804,7 @@ fn handle_safekeeper(sub_match: &ArgMatches, env: &local_env::LocalEnv) -> Resul
 }

 fn handle_start_all(sub_match: &ArgMatches, env: &local_env::LocalEnv) -> anyhow::Result<()> {
-    // Endpoints are not started automatically
+    // Postgres nodes are not started automatically

    broker::start_broker_process(env)?;

@@ -843,10 +838,10 @@ fn handle_stop_all(sub_match: &ArgMatches, env: &local_env::LocalEnv) -> Result<
 fn try_stop_all(env: &local_env::LocalEnv, immediate: bool) {
    let pageserver = PageServerNode::from_env(env);

-    // Stop all endpoints
+    // Stop all compute nodes
    match ComputeControlPlane::load(env.clone()) {
        Ok(cplane) => {
-            for (_k, node) in cplane.endpoints {
+            for (_k, node) in cplane.nodes {
                if let Err(e) = node.stop(false) {
                    eprintln!("postgres stop failed: {e:#}");
                }
@@ -879,9 +874,7 @@ fn cli() -> Command {
        .help("Name of the branch to be created or used as an alias for other services")
        .required(false);

-    let endpoint_id_arg = Arg::new("endpoint_id")
-        .help("Postgres endpoint id")
-        .required(false);
+    let pg_node_arg = Arg::new("node").help("Postgres node name").required(false);

    let safekeeper_id_arg = Arg::new("id").help("safekeeper id").required(false);

@@ -1035,27 +1028,27 @@ fn cli() -> Command {
                )
        )
        .subcommand(
-            Command::new("endpoint")
+            Command::new("pg")
                .arg_required_else_help(true)
                .about("Manage postgres instances")
                .subcommand(Command::new("list").arg(tenant_id_arg.clone()))
                .subcommand(Command::new("create")
-                    .about("Create a compute endpoint")
-                    .arg(endpoint_id_arg.clone())
+                    .about("Create a postgres compute node")
+                    .arg(pg_node_arg.clone())
                    .arg(branch_name_arg.clone())
                    .arg(tenant_id_arg.clone())
                    .arg(lsn_arg.clone())
                    .arg(port_arg.clone())
                    .arg(
                        Arg::new("config-only")
-                            .help("Don't do basebackup, create endpoint directory with only config files")
+                            .help("Don't do basebackup, create compute node with only config files")
                            .long("config-only")
                            .required(false))
                    .arg(pg_version_arg.clone())
                )
                .subcommand(Command::new("start")
-                    .about("Start postgres.\n If the endpoint doesn't exist yet, it is created.")
-                    .arg(endpoint_id_arg.clone())
+                    .about("Start a postgres compute node.\n This command actually creates new node from scratch, but preserves existing config files")
+                    .arg(pg_node_arg.clone())
                    .arg(tenant_id_arg.clone())
                    .arg(branch_name_arg)
                    .arg(timeline_id_arg)
@@ -1065,7 +1058,7 @@ fn cli() -> Command {
                )
                .subcommand(
                    Command::new("stop")
-                    .arg(endpoint_id_arg)
+                    .arg(pg_node_arg)
                    .arg(tenant_id_arg)
                    .arg(
                        Arg::new("destroy")
@@ -1077,13 +1070,6 @@ fn cli() -> Command {
                )

        )
-        // Obsolete old name for 'endpoint'. We now just print an error if it's used.
-        .subcommand(
-            Command::new("pg")
-                .hide(true)
-                .arg(Arg::new("ignore-rest").allow_hyphen_values(true).num_args(0..).required(false))
-                .trailing_var_arg(true)
-        )
        .subcommand(
            Command::new("start")
                .about("Start page server and safekeepers")
--- a/control_plane/src/endpoint.rs
+++ b/control_plane/src/endpoint.rs
@@ -25,45 +25,54 @@ use crate::postgresql_conf::PostgresConf;
 //
 pub struct ComputeControlPlane {
    base_port: u16,
-
-    // endpoint ID is the key
-    pub endpoints: BTreeMap<String, Arc<Endpoint>>,
-
-    env: LocalEnv,
    pageserver: Arc<PageServerNode>,
+    pub nodes: BTreeMap<(TenantId, String), Arc<PostgresNode>>,
+    env: LocalEnv,
 }

 impl ComputeControlPlane {
-    // Load current endpoints from the endpoints/ subdirectories
+    // Load current nodes with ports from data directories on disk
+    // Directory structure has the following layout:
+    // pgdatadirs
+    // |- tenants
+    // |  |- <tenant_id>
+    // |  |   |- <node name>
    pub fn load(env: LocalEnv) -> Result<ComputeControlPlane> {
        let pageserver = Arc::new(PageServerNode::from_env(&env));

-        let mut endpoints = BTreeMap::default();
-        for endpoint_dir in fs::read_dir(env.endpoints_path())
-            .with_context(|| format!("failed to list {}", env.endpoints_path().display()))?
+        let mut nodes = BTreeMap::default();
+        let pgdatadirspath = &env.pg_data_dirs_path();
+
+        for tenant_dir in fs::read_dir(pgdatadirspath)
+            .with_context(|| format!("failed to list {}", pgdatadirspath.display()))?
        {
-            let ep = Endpoint::from_dir_entry(endpoint_dir?, &env, &pageserver)?;
-            endpoints.insert(ep.name.clone(), Arc::new(ep));
+            let tenant_dir = tenant_dir?;
+            for timeline_dir in fs::read_dir(tenant_dir.path())
+                .with_context(|| format!("failed to list {}", tenant_dir.path().display()))?
+            {
+                let node = PostgresNode::from_dir_entry(timeline_dir?, &env, &pageserver)?;
+                nodes.insert((node.tenant_id, node.name.clone()), Arc::new(node));
+            }
        }

        Ok(ComputeControlPlane {
            base_port: 55431,
-            endpoints,
-            env,
            pageserver,
+            nodes,
+            env,
        })
    }

    fn get_port(&mut self) -> u16 {
        1 + self
-            .endpoints
+            .nodes
            .values()
-            .map(|ep| ep.address.port())
+            .map(|node| node.address.port())
            .max()
            .unwrap_or(self.base_port)
    }

-    pub fn new_endpoint(
+    pub fn new_node(
        &mut self,
        tenant_id: TenantId,
        name: &str,
@@ -71,9 +80,9 @@ impl ComputeControlPlane {
        lsn: Option<Lsn>,
        port: Option<u16>,
        pg_version: u32,
-    ) -> Result<Arc<Endpoint>> {
+    ) -> Result<Arc<PostgresNode>> {
        let port = port.unwrap_or_else(|| self.get_port());
-        let ep = Arc::new(Endpoint {
+        let node = Arc::new(PostgresNode {
            name: name.to_owned(),
            address: SocketAddr::new("127.0.0.1".parse().unwrap(), port),
            env: self.env.clone(),
@@ -81,48 +90,44 @@ impl ComputeControlPlane {
            timeline_id,
            lsn,
            tenant_id,
+            uses_wal_proposer: false,
            pg_version,
        });

-        ep.create_pgdata()?;
-        ep.setup_pg_conf()?;
+        node.create_pgdata()?;
+        node.setup_pg_conf()?;

-        self.endpoints.insert(ep.name.clone(), Arc::clone(&ep));
+        self.nodes
+            .insert((tenant_id, node.name.clone()), Arc::clone(&node));

-        Ok(ep)
+        Ok(node)
    }
 }

 ///////////////////////////////////////////////////////////////////////////////

 #[derive(Debug)]
-pub struct Endpoint {
-    /// used as the directory name
-    name: String,
-    pub tenant_id: TenantId,
-    pub timeline_id: TimelineId,
-    // Some(lsn) if this is a read-only endpoint anchored at 'lsn'. None for the primary.
-    pub lsn: Option<Lsn>,
-
-    // port and address of the Postgres server
+pub struct PostgresNode {
    pub address: SocketAddr,
-    pg_version: u32,
-
-    // These are not part of the endpoint as such, but the environment
-    // the endpoint runs in.
+    name: String,
    pub env: LocalEnv,
    pageserver: Arc<PageServerNode>,
+    pub timeline_id: TimelineId,
+    pub lsn: Option<Lsn>, // if it's a read-only node. None for primary
+    pub tenant_id: TenantId,
+    uses_wal_proposer: bool,
+    pg_version: u32,
 }

-impl Endpoint {
+impl PostgresNode {
    fn from_dir_entry(
        entry: std::fs::DirEntry,
        env: &LocalEnv,
        pageserver: &Arc<PageServerNode>,
-    ) -> Result<Endpoint> {
+    ) -> Result<PostgresNode> {
        if !entry.file_type()?.is_dir() {
            anyhow::bail!(
-                "Endpoint::from_dir_entry failed: '{}' is not a directory",
+                "PostgresNode::from_dir_entry failed: '{}' is not a directory",
                entry.path().display()
            );
        }
@@ -132,7 +137,7 @@ impl Endpoint {
        let name = fname.to_str().unwrap().to_string();

        // Read config file into memory
-        let cfg_path = entry.path().join("pgdata").join("postgresql.conf");
+        let cfg_path = entry.path().join("postgresql.conf");
        let cfg_path_str = cfg_path.to_string_lossy();
        let mut conf_file = File::open(&cfg_path)
            .with_context(|| format!("failed to open config file in {}", cfg_path_str))?;
@@ -144,6 +149,7 @@ impl Endpoint {
        let port: u16 = conf.parse_field("port", &context)?;
        let timeline_id: TimelineId = conf.parse_field("neon.timeline_id", &context)?;
        let tenant_id: TenantId = conf.parse_field("neon.tenant_id", &context)?;
+        let uses_wal_proposer = conf.get("neon.safekeepers").is_some();

        // Read postgres version from PG_VERSION file to determine which postgres version binary to use.
        // If it doesn't exist, assume broken data directory and use default pg version.
@@ -158,7 +164,7 @@ impl Endpoint {
            conf.parse_field_optional("recovery_target_lsn", &context)?;

        // ok now
-        Ok(Endpoint {
+        Ok(PostgresNode {
            address: SocketAddr::new("127.0.0.1".parse().unwrap(), port),
            name,
            env: env.clone(),
@@ -166,6 +172,7 @@ impl Endpoint {
            timeline_id,
            lsn: recovery_target_lsn,
            tenant_id,
+            uses_wal_proposer,
            pg_version,
        })
    }
@@ -266,7 +273,7 @@ impl Endpoint {
    }

    // Write postgresql.conf with default configuration
-    // and PG_VERSION file to the data directory of a new endpoint.
+    // and PG_VERSION file to the data directory of a new node.
    fn setup_pg_conf(&self) -> Result<()> {
        let mut conf = PostgresConf::new();
        conf.append("max_wal_senders", "10");
@@ -286,7 +293,7 @@ impl Endpoint {
        // walproposer panics when basebackup is invalid, it is pointless to restart in this case.
        conf.append("restart_after_crash", "off");

-        // Configure the Neon Postgres extension to fetch pages from pageserver
+        // Configure the node to fetch pages from pageserver
        let pageserver_connstr = {
            let config = &self.pageserver.pg_connection_config;
            let (host, port) = (config.host(), config.port());
@@ -322,7 +329,7 @@ impl Endpoint {
        conf.append("max_replication_flush_lag", "10GB");

        if !self.env.safekeepers.is_empty() {
-            // Configure Postgres to connect to the safekeepers
+            // Configure the node to connect to the safekeepers
            conf.append("synchronous_standby_names", "walproposer");

            let safekeepers = self
@@ -357,7 +364,7 @@ impl Endpoint {
    fn load_basebackup(&self, auth_token: &Option<String>) -> Result<()> {
        let backup_lsn = if let Some(lsn) = self.lsn {
            Some(lsn)
-        } else if !self.env.safekeepers.is_empty() {
+        } else if self.uses_wal_proposer {
            // LSN 0 means that it is bootstrap and we need to download just
            // latest data from the pageserver. That is a bit clumsy but whole bootstrap
            // procedure evolves quite actively right now, so let's think about it again
@@ -377,12 +384,8 @@ impl Endpoint {
        Ok(())
    }

-    pub fn endpoint_path(&self) -> PathBuf {
-        self.env.endpoints_path().join(&self.name)
-    }
-
    pub fn pgdata(&self) -> PathBuf {
-        self.endpoint_path().join("pgdata")
+        self.env.pg_data_dir(&self.tenant_id, &self.name)
    }

    pub fn status(&self) -> &str {
@@ -444,11 +447,12 @@ impl Endpoint {
    }

    pub fn start(&self, auth_token: &Option<String>) -> Result<()> {
+        // Bail if the node already running.
        if self.status() == "running" {
-            anyhow::bail!("The endpoint is already running");
+            anyhow::bail!("The node is already running");
        }

-        // 1. We always start Postgres from scratch, so
+        // 1. We always start compute node from scratch, so
        // if old dir exists, preserve 'postgresql.conf' and drop the directory
        let postgresql_conf_path = self.pgdata().join("postgresql.conf");
        let postgresql_conf = fs::read(&postgresql_conf_path).with_context(|| {
@@ -470,8 +474,8 @@ impl Endpoint {
            File::create(self.pgdata().join("standby.signal"))?;
        }

-        // 4. Finally start postgres
-        println!("Starting postgres at '{}'", self.connstr());
+        // 4. Finally start the compute node postgres
+        println!("Starting postgres node at '{}'", self.connstr());
        self.pg_ctl(&["start"], auth_token)
    }

@@ -480,7 +484,7 @@ impl Endpoint {
        // use immediate shutdown mode, otherwise,
        // shutdown gracefully to leave the data directory sane.
        //
-        // Postgres is always started from scratch, so stop
+        // Compute node always starts from scratch, so stop
        // without destroy only used for testing and debugging.
        //
        if destroy {
@@ -489,7 +493,7 @@ impl Endpoint {
                "Destroying postgres data directory '{}'",
                self.pgdata().to_str().unwrap()
            );
-            fs::remove_dir_all(self.endpoint_path())?;
+            fs::remove_dir_all(self.pgdata())?;
        } else {
            self.pg_ctl(&["stop"], &None)?;
        }
--- a/control_plane/src/lib.rs
+++ b/control_plane/src/lib.rs
@@ -9,7 +9,7 @@

 mod background_process;
 pub mod broker;
-pub mod endpoint;
+pub mod compute;
 pub mod local_env;
 pub mod pageserver;
 pub mod postgresql_conf;
--- a/control_plane/src/local_env.rs
+++ b/control_plane/src/local_env.rs
@@ -200,8 +200,14 @@ impl LocalEnv {
        self.neon_distrib_dir.join("storage_broker")
    }

-    pub fn endpoints_path(&self) -> PathBuf {
-        self.base_data_dir.join("endpoints")
+    pub fn pg_data_dirs_path(&self) -> PathBuf {
+        self.base_data_dir.join("pgdatadirs").join("tenants")
+    }
+
+    pub fn pg_data_dir(&self, tenant_id: &TenantId, branch_name: &str) -> PathBuf {
+        self.pg_data_dirs_path()
+            .join(tenant_id.to_string())
+            .join(branch_name)
    }

    // TODO: move pageserver files into ./pageserver
@@ -421,7 +427,7 @@ impl LocalEnv {
            }
        }

-        fs::create_dir_all(self.endpoints_path())?;
+        fs::create_dir_all(self.pg_data_dirs_path())?;

        for safekeeper in &self.safekeepers {
            fs::create_dir_all(SafekeeperNode::datadir_path_by_id(self, safekeeper.id))?;
--- a/control_plane/src/pageserver.rs
+++ b/control_plane/src/pageserver.rs
@@ -363,11 +363,6 @@ impl PageServerNode {
                .map(|x| serde_json::from_str(x))
                .transpose()
                .context("Failed to parse 'eviction_policy' json")?,
-            min_resident_size_override: settings
-                .remove("min_resident_size_override")
-                .map(|x| x.parse::<u64>())
-                .transpose()
-                .context("Failed to parse 'min_resident_size_override' as integer")?,
        };
        if !settings.is_empty() {
            bail!("Unrecognized tenant settings: {settings:?}")
@@ -440,11 +435,6 @@ impl PageServerNode {
                    .map(|x| serde_json::from_str(x))
                    .transpose()
                    .context("Failed to parse 'eviction_policy' json")?,
-                min_resident_size_override: settings
-                    .get("min_resident_size_override")
-                    .map(|x| x.parse::<u64>())
-                    .transpose()
-                    .context("Failed to parse 'min_resident_size_override' as an integer")?,
            })
            .send()?
            .error_from_body()?;
--- a/libs/compute_api/Cargo.toml
+++ b/libs/compute_api/Cargo.toml
@@ -1,14 +0,0 @@
-[package]
-name = "compute_api"
-version = "0.1.0"
-edition.workspace = true
-license.workspace = true
-
-[dependencies]
-anyhow.workspace = true
-chrono.workspace = true
-serde.workspace = true
-serde_with.workspace = true
-serde_json.workspace = true
-
-workspace_hack.workspace = true
--- a/libs/compute_api/src/lib.rs
+++ b/libs/compute_api/src/lib.rs
@@ -1,3 +0,0 @@
-pub mod requests;
-pub mod responses;
-pub mod spec;
--- a/libs/compute_api/src/requests.rs
+++ b/libs/compute_api/src/requests.rs
@@ -1,14 +0,0 @@
-//! Structs representing the JSON formats used in the compute_ctl's HTTP API.
-
-use crate::spec::ComputeSpec;
-use serde::Deserialize;
-
-/// Request of the /configure API
-///
-/// We now pass only `spec` in the configuration request, but later we can
-/// extend it and something like `restart: bool` or something else. So put
-/// `spec` into a struct initially to be more flexible in the future.
-#[derive(Deserialize, Debug)]
-pub struct ConfigurationRequest {
-    pub spec: ComputeSpec,
-}
--- a/libs/compute_api/src/responses.rs
+++ b/libs/compute_api/src/responses.rs
@@ -1,78 +0,0 @@
-//! Structs representing the JSON formats used in the compute_ctl's HTTP API.
-
-use chrono::{DateTime, Utc};
-use serde::{Deserialize, Serialize, Serializer};
-
-use crate::spec::ComputeSpec;
-
-#[derive(Serialize, Debug)]
-pub struct GenericAPIError {
-    pub error: String,
-}
-
-/// Response of the /status API
-#[derive(Serialize, Debug)]
-#[serde(rename_all = "snake_case")]
-pub struct ComputeStatusResponse {
-    pub tenant: Option<String>,
-    pub timeline: Option<String>,
-    pub status: ComputeStatus,
-    #[serde(serialize_with = "rfc3339_serialize")]
-    pub last_active: DateTime<Utc>,
-    pub error: Option<String>,
-}
-
-#[derive(Serialize)]
-#[serde(rename_all = "snake_case")]
-pub struct ComputeState {
-    pub status: ComputeStatus,
-    /// Timestamp of the last Postgres activity
-    #[serde(serialize_with = "rfc3339_serialize")]
-    pub last_active: DateTime<Utc>,
-    pub error: Option<String>,
-}
-
-#[derive(Serialize, Clone, Copy, Debug, PartialEq, Eq)]
-#[serde(rename_all = "snake_case")]
-pub enum ComputeStatus {
-    // Spec wasn't provided at start, waiting for it to be
-    // provided by control-plane.
-    Empty,
-    // Compute configuration was requested.
-    ConfigurationPending,
-    // Compute node has spec and initial startup and
-    // configuration is in progress.
-    Init,
-    // Compute is configured and running.
-    Running,
-    // New spec is being applied.
-    Configuration,
-    // Either startup or configuration failed,
-    // compute will exit soon or is waiting for
-    // control-plane to terminate it.
-    Failed,
-}
-
-fn rfc3339_serialize<S>(x: &DateTime<Utc>, s: S) -> Result<S::Ok, S::Error>
-where
-    S: Serializer,
-{
-    x.to_rfc3339().serialize(s)
-}
-
-/// Response of the /metrics.json API
-#[derive(Clone, Debug, Default, Serialize)]
-pub struct ComputeMetrics {
-    pub sync_safekeepers_ms: u64,
-    pub basebackup_ms: u64,
-    pub config_ms: u64,
-    pub total_startup_ms: u64,
-}
-
-/// Response of the `/computes/{compute_id}/spec` control-plane API.
-/// This is not actually a compute API response, so consider moving
-/// to a different place.
-#[derive(Deserialize, Debug)]
-pub struct ControlPlaneSpecResponse {
-    pub spec: Option<ComputeSpec>,
-}
--- a/libs/compute_api/src/spec.rs
+++ b/libs/compute_api/src/spec.rs
@@ -1,97 +0,0 @@
-//! `ComputeSpec` represents the contents of the spec.json file.
-//!
-//! The spec.json file is used to pass information to 'compute_ctl'. It contains
-//! all the information needed to start up the right version of PostgreSQL,
-//! and connect it to the storage nodes.
-use serde::Deserialize;
-use std::collections::HashMap;
-
-/// String type alias representing Postgres identifier and
-/// intended to be used for DB / role names.
-pub type PgIdent = String;
-
-/// Cluster spec or configuration represented as an optional number of
-/// delta operations + final cluster state description.
-#[derive(Clone, Debug, Default, Deserialize)]
-pub struct ComputeSpec {
-    pub format_version: f32,
-
-    // The control plane also includes a 'timestamp' field in the JSON document,
-    // but we don't use it for anything. Serde will ignore missing fields when
-    // deserializing it.
-    pub operation_uuid: Option<String>,
-    /// Expected cluster state at the end of transition process.
-    pub cluster: Cluster,
-    pub delta_operations: Option<Vec<DeltaOp>>,
-
-    pub storage_auth_token: Option<String>,
-
-    pub startup_tracing_context: Option<HashMap<String, String>>,
-}
-
-#[derive(Clone, Debug, Default, Deserialize)]
-pub struct Cluster {
-    pub cluster_id: String,
-    pub name: String,
-    pub state: Option<String>,
-    pub roles: Vec<Role>,
-    pub databases: Vec<Database>,
-    pub settings: GenericOptions,
-}
-
-/// Single cluster state changing operation that could not be represented as
-/// a static `Cluster` structure. For example:
-/// - DROP DATABASE
-/// - DROP ROLE
-/// - ALTER ROLE name RENAME TO new_name
-/// - ALTER DATABASE name RENAME TO new_name
-#[derive(Clone, Debug, Deserialize)]
-pub struct DeltaOp {
-    pub action: String,
-    pub name: PgIdent,
-    pub new_name: Option<PgIdent>,
-}
-
-/// Rust representation of Postgres role info with only those fields
-/// that matter for us.
-#[derive(Clone, Debug, Deserialize)]
-pub struct Role {
-    pub name: PgIdent,
-    pub encrypted_password: Option<String>,
-    pub options: GenericOptions,
-}
-
-/// Rust representation of Postgres database info with only those fields
-/// that matter for us.
-#[derive(Clone, Debug, Deserialize)]
-pub struct Database {
-    pub name: PgIdent,
-    pub owner: PgIdent,
-    pub options: GenericOptions,
-}
-
-/// Common type representing both SQL statement params with or without value,
-/// like `LOGIN` or `OWNER username` in the `CREATE/ALTER ROLE`, and config
-/// options like `wal_level = logical`.
-#[derive(Clone, Debug, Deserialize)]
-pub struct GenericOption {
-    pub name: String,
-    pub value: Option<String>,
-    pub vartype: String,
-}
-
-/// Optional collection of `GenericOption`'s. Type alias allows us to
-/// declare a `trait` on it.
-pub type GenericOptions = Option<Vec<GenericOption>>;
-
-#[cfg(test)]
-mod tests {
-    use super::*;
-    use std::fs::File;
-
-    #[test]
-    fn parse_spec_file() {
-        let file = File::open("tests/cluster_spec.json").unwrap();
-        let _spec: ComputeSpec = serde_json::from_reader(file).unwrap();
-    }
-}
--- a/libs/pageserver_api/Cargo.toml
+++ b/libs/pageserver_api/Cargo.toml
@@ -7,7 +7,6 @@ license.workspace = true
 [dependencies]
 serde.workspace = true
 serde_with.workspace = true
-serde_json.workspace = true
 const_format.workspace = true
 anyhow.workspace = true
 bytes.workspace = true
@@ -15,7 +14,6 @@ byteorder.workspace = true
 utils.workspace = true
 postgres_ffi.workspace = true
 enum-map.workspace = true
-strum.workspace = true
-strum_macros.workspace = true
+serde_json.workspace = true

 workspace_hack.workspace = true
--- a/libs/pageserver_api/src/models.rs
+++ b/libs/pageserver_api/src/models.rs
@@ -7,7 +7,6 @@ use std::{
 use byteorder::{BigEndian, ReadBytesExt};
 use serde::{Deserialize, Serialize};
 use serde_with::{serde_as, DisplayFromStr};
-use strum_macros;
 use utils::{
    history_buffer::HistoryBufferWithDropCounter,
    id::{NodeId, TenantId, TimelineId},
@@ -19,23 +18,11 @@ use anyhow::bail;
 use bytes::{BufMut, Bytes, BytesMut};

 /// A state of a tenant in pageserver's memory.
-#[derive(
-    Clone,
-    PartialEq,
-    Eq,
-    serde::Serialize,
-    serde::Deserialize,
-    strum_macros::Display,
-    strum_macros::EnumString,
-    strum_macros::EnumVariantNames,
-    strum_macros::AsRefStr,
-    strum_macros::IntoStaticStr,
-)]
-#[serde(tag = "slug", content = "data")]
+#[derive(Debug, Clone, Copy, PartialEq, Eq, serde::Serialize, serde::Deserialize)]
 pub enum TenantState {
-    /// This tenant is being loaded from local disk
+    // This tenant is being loaded from local disk
    Loading,
-    /// This tenant is being downloaded from cloud storage.
+    // This tenant is being downloaded from cloud storage.
    Attaching,
    /// Tenant is fully operational
    Active,
@@ -44,7 +31,15 @@ pub enum TenantState {
    Stopping,
    /// A tenant is recognized by the pageserver, but can no longer be used for
    /// any operations, because it failed to be activated.
-    Broken { reason: String, backtrace: String },
+    Broken,
+}
+
+pub mod state {
+    pub const LOADING: &str = "loading";
+    pub const ATTACHING: &str = "attaching";
+    pub const ACTIVE: &str = "active";
+    pub const STOPPING: &str = "stopping";
+    pub const BROKEN: &str = "broken";
 }

 impl TenantState {
@@ -54,26 +49,17 @@ impl TenantState {
            Self::Attaching => true,
            Self::Active => false,
            Self::Stopping => false,
-            Self::Broken { .. } => false,
+            Self::Broken => false,
        }
    }

-    pub fn broken_from_reason(reason: String) -> Self {
-        let backtrace_str: String = format!("{}", std::backtrace::Backtrace::force_capture());
-        Self::Broken {
-            reason,
-            backtrace: backtrace_str,
-        }
-    }
-}
-
-impl std::fmt::Debug for TenantState {
-    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+    pub fn as_str(&self) -> &'static str {
        match self {
-            Self::Broken { reason, backtrace } if !reason.is_empty() => {
-                write!(f, "Broken due to: {reason}. Backtrace:\n{backtrace}")
-            }
-            _ => write!(f, "{self}"),
+            TenantState::Loading => state::LOADING,
+            TenantState::Attaching => state::ATTACHING,
+            TenantState::Active => state::ACTIVE,
+            TenantState::Stopping => state::STOPPING,
+            TenantState::Broken => state::BROKEN,
        }
    }
 }
@@ -134,7 +120,6 @@ pub struct TenantCreateRequest {
    // We might do that once the eviction feature has stabilizied.
    // For now, this field is not even documented in the openapi_spec.yml.
    pub eviction_policy: Option<serde_json::Value>,
-    pub min_resident_size_override: Option<u64>,
 }

 #[serde_as]
@@ -180,7 +165,6 @@ pub struct TenantConfigRequest {
    // We might do that once the eviction feature has stabilizied.
    // For now, this field is not even documented in the openapi_spec.yml.
    pub eviction_policy: Option<serde_json::Value>,
-    pub min_resident_size_override: Option<u64>,
 }

 impl TenantConfigRequest {
@@ -201,7 +185,6 @@ impl TenantConfigRequest {
            max_lsn_wal_lag: None,
            trace_read_requests: None,
            eviction_policy: None,
-            min_resident_size_override: None,
        }
    }
 }
@@ -629,7 +612,6 @@ impl PagestreamBeMessage {
 #[cfg(test)]
 mod tests {
    use bytes::Buf;
-    use serde_json::json;

    use super::*;

@@ -680,57 +662,4 @@ mod tests {
            assert!(msg == reconstructed);
        }
    }
-
-    #[test]
-    fn test_tenantinfo_serde() {
-        // Test serialization/deserialization of TenantInfo
-        let original_active = TenantInfo {
-            id: TenantId::generate(),
-            state: TenantState::Active,
-            current_physical_size: Some(42),
-            has_in_progress_downloads: Some(false),
-        };
-        let expected_active = json!({
-            "id": original_active.id.to_string(),
-            "state": {
-                "slug": "Active",
-            },
-            "current_physical_size": 42,
-            "has_in_progress_downloads": false,
-        });
-
-        let original_broken = TenantInfo {
-            id: TenantId::generate(),
-            state: TenantState::Broken {
-                reason: "reason".into(),
-                backtrace: "backtrace info".into(),
-            },
-            current_physical_size: Some(42),
-            has_in_progress_downloads: Some(false),
-        };
-        let expected_broken = json!({
-            "id": original_broken.id.to_string(),
-            "state": {
-                "slug": "Broken",
-                "data": {
-                    "backtrace": "backtrace info",
-                    "reason": "reason",
-                }
-            },
-            "current_physical_size": 42,
-            "has_in_progress_downloads": false,
-        });
-
-        assert_eq!(
-            serde_json::to_value(&original_active).unwrap(),
-            expected_active
-        );
-
-        assert_eq!(
-            serde_json::to_value(&original_broken).unwrap(),
-            expected_broken
-        );
-        assert!(format!("{:?}", &original_broken.state).contains("reason"));
-        assert!(format!("{:?}", &original_broken.state).contains("backtrace info"));
-    }
 }
--- a/libs/postgres_backend/src/lib.rs
+++ b/libs/postgres_backend/src/lib.rs
@@ -54,7 +54,7 @@ pub fn is_expected_io_error(e: &io::Error) -> bool {
    use io::ErrorKind::*;
    matches!(
        e.kind(),
-        ConnectionRefused | ConnectionAborted | ConnectionReset | TimedOut
+        ConnectionRefused | ConnectionAborted | ConnectionReset
    )
 }

@@ -320,17 +320,9 @@ impl<IO: AsyncRead + AsyncWrite + Unpin> PostgresBackend<IO> {
        if let ProtoState::Closed = self.state {
            Ok(None)
        } else {
-            match self.framed.read_message().await {
-                Ok(m) => {
-                    trace!("read msg {:?}", m);
-                    Ok(m)
-                }
-                Err(e) => {
-                    // remember not to try to read anymore
-                    self.state = ProtoState::Closed;
-                    Err(e)
-                }
-            }
+            let m = self.framed.read_message().await?;
+            trace!("read msg {:?}", m);
+            Ok(m)
        }
    }

@@ -501,10 +493,7 @@ impl<IO: AsyncRead + AsyncWrite + Unpin> PostgresBackend<IO> {
            MaybeWriteOnly::Full(framed) => {
                let (reader, writer) = framed.split();
                self.framed = MaybeWriteOnly::WriteOnly(writer);
-                Ok(PostgresBackendReader {
-                    reader,
-                    closed: false,
-                })
+                Ok(PostgresBackendReader(reader))
            }
            MaybeWriteOnly::WriteOnly(_) => {
                anyhow::bail!("PostgresBackend is already split")
@@ -521,12 +510,8 @@ impl<IO: AsyncRead + AsyncWrite + Unpin> PostgresBackend<IO> {
                anyhow::bail!("PostgresBackend is not split")
            }
            MaybeWriteOnly::WriteOnly(writer) => {
-                let joined = Framed::unsplit(reader.reader, writer);
+                let joined = Framed::unsplit(reader.0, writer);
                self.framed = MaybeWriteOnly::Full(joined);
-                // if reader encountered connection error, do not attempt reading anymore
-                if reader.closed {
-                    self.state = ProtoState::Closed;
-                }
                Ok(())
            }
            MaybeWriteOnly::Broken => panic!("unsplit on framed in invalid state"),
@@ -812,25 +797,15 @@ impl<IO: AsyncRead + AsyncWrite + Unpin> PostgresBackend<IO> {
    }
 }

-pub struct PostgresBackendReader<IO> {
-    reader: FramedReader<MaybeTlsStream<IO>>,
-    closed: bool, // true if received error closing the connection
-}
+pub struct PostgresBackendReader<IO>(FramedReader<MaybeTlsStream<IO>>);

 impl<IO: AsyncRead + AsyncWrite + Unpin> PostgresBackendReader<IO> {
    /// Read full message or return None if connection is cleanly closed with no
    /// unprocessed data.
    pub async fn read_message(&mut self) -> Result<Option<FeMessage>, ConnectionError> {
-        match self.reader.read_message().await {
-            Ok(m) => {
-                trace!("read msg {:?}", m);
-                Ok(m)
-            }
-            Err(e) => {
-                self.closed = true;
-                Err(e)
-            }
-        }
+        let m = self.0.read_message().await?;
+        trace!("read msg {:?}", m);
+        Ok(m)
    }

    /// Get CopyData contents of the next message in COPY stream or error
@@ -948,7 +923,7 @@ pub enum CopyStreamHandlerEnd {
    #[error("EOF on COPY stream")]
    EOF,
    /// The connection was lost
-    #[error("connection error: {0}")]
+    #[error(transparent)]
    Disconnected(#[from] ConnectionError),
    /// Some other error
    #[error(transparent)]
--- a/libs/pq_proto/src/lib.rs
+++ b/libs/pq_proto/src/lib.rs
@@ -293,9 +293,6 @@ impl FeStartupPacket {
        // We shouldn't advance `buf` as probably full message is not there yet,
        // so can't directly use Bytes::get_u32 etc.
        let len = (&buf[0..4]).read_u32::<BigEndian>().unwrap() as usize;
-        // The proposed replacement is `!(4..=MAX_STARTUP_PACKET_LENGTH).contains(&len)`
-        // which is less readable
-        #[allow(clippy::manual_range_contains)]
        if len < 4 || len > MAX_STARTUP_PACKET_LENGTH {
            return Err(ProtocolError::Protocol(format!(
                "invalid startup packet message length {}",
@@ -939,40 +936,35 @@ impl<'a> BeMessage<'a> {
    }
 }

-/// Feedback pageserver sends to safekeeper and safekeeper resends to compute.
-/// Serialized in custom flexible key/value format. In replication protocol, it
-/// is marked with NEON_STATUS_UPDATE_TAG_BYTE to differentiate from postgres
-/// Standby status update / Hot standby feedback messages.
+// Neon extension of postgres replication protocol
+// See NEON_STATUS_UPDATE_TAG_BYTE
 #[derive(Debug, Clone, Copy, PartialEq, Eq, Serialize, Deserialize)]
-pub struct PageserverFeedback {
-    /// Last known size of the timeline. Used to enforce timeline size limit.
+pub struct ReplicationFeedback {
+    // Last known size of the timeline. Used to enforce timeline size limit.
    pub current_timeline_size: u64,
-    /// LSN last received and ingested by the pageserver.
-    pub last_received_lsn: u64,
-    /// LSN up to which data is persisted by the pageserver to its local disc.
-    pub disk_consistent_lsn: u64,
-    /// LSN up to which data is persisted by the pageserver on s3; safekeepers
-    /// consider WAL before it can be removed.
-    pub remote_consistent_lsn: u64,
-    pub replytime: SystemTime,
+    // Parts of StandbyStatusUpdate we resend to compute via safekeeper
+    pub ps_writelsn: u64,
+    pub ps_applylsn: u64,
+    pub ps_flushlsn: u64,
+    pub ps_replytime: SystemTime,
 }

-// NOTE: Do not forget to increment this number when adding new fields to PageserverFeedback.
+// NOTE: Do not forget to increment this number when adding new fields to ReplicationFeedback.
 // Do not remove previously available fields because this might be backwards incompatible.
-pub const PAGESERVER_FEEDBACK_FIELDS_NUMBER: u8 = 5;
+pub const REPLICATION_FEEDBACK_FIELDS_NUMBER: u8 = 5;

-impl PageserverFeedback {
-    pub fn empty() -> PageserverFeedback {
-        PageserverFeedback {
+impl ReplicationFeedback {
+    pub fn empty() -> ReplicationFeedback {
+        ReplicationFeedback {
            current_timeline_size: 0,
-            last_received_lsn: 0,
-            remote_consistent_lsn: 0,
-            disk_consistent_lsn: 0,
-            replytime: SystemTime::now(),
+            ps_writelsn: 0,
+            ps_applylsn: 0,
+            ps_flushlsn: 0,
+            ps_replytime: SystemTime::now(),
        }
    }

-    // Serialize PageserverFeedback using custom format
+    // Serialize ReplicationFeedback using custom format
    // to support protocol extensibility.
    //
    // Following layout is used:
@@ -982,26 +974,24 @@ impl PageserverFeedback {
    // null-terminated string - key,
    // uint32 - value length in bytes
    // value itself
-    //
-    // TODO: change serialized fields names once all computes migrate to rename.
    pub fn serialize(&self, buf: &mut BytesMut) {
-        buf.put_u8(PAGESERVER_FEEDBACK_FIELDS_NUMBER); // # of keys
+        buf.put_u8(REPLICATION_FEEDBACK_FIELDS_NUMBER); // # of keys
        buf.put_slice(b"current_timeline_size\0");
        buf.put_i32(8);
        buf.put_u64(self.current_timeline_size);

        buf.put_slice(b"ps_writelsn\0");
        buf.put_i32(8);
-        buf.put_u64(self.last_received_lsn);
+        buf.put_u64(self.ps_writelsn);
        buf.put_slice(b"ps_flushlsn\0");
        buf.put_i32(8);
-        buf.put_u64(self.disk_consistent_lsn);
+        buf.put_u64(self.ps_flushlsn);
        buf.put_slice(b"ps_applylsn\0");
        buf.put_i32(8);
-        buf.put_u64(self.remote_consistent_lsn);
+        buf.put_u64(self.ps_applylsn);

        let timestamp = self
-            .replytime
+            .ps_replytime
            .duration_since(*PG_EPOCH)
            .expect("failed to serialize pg_replytime earlier than PG_EPOCH")
            .as_micros() as i64;
@@ -1011,10 +1001,9 @@ impl PageserverFeedback {
        buf.put_i64(timestamp);
    }

-    // Deserialize PageserverFeedback message
-    // TODO: change serialized fields names once all computes migrate to rename.
-    pub fn parse(mut buf: Bytes) -> PageserverFeedback {
-        let mut rf = PageserverFeedback::empty();
+    // Deserialize ReplicationFeedback message
+    pub fn parse(mut buf: Bytes) -> ReplicationFeedback {
+        let mut rf = ReplicationFeedback::empty();
        let nfields = buf.get_u8();
        for _ in 0..nfields {
            let key = read_cstr(&mut buf).unwrap();
@@ -1027,39 +1016,39 @@ impl PageserverFeedback {
                b"ps_writelsn" => {
                    let len = buf.get_i32();
                    assert_eq!(len, 8);
-                    rf.last_received_lsn = buf.get_u64();
+                    rf.ps_writelsn = buf.get_u64();
                }
                b"ps_flushlsn" => {
                    let len = buf.get_i32();
                    assert_eq!(len, 8);
-                    rf.disk_consistent_lsn = buf.get_u64();
+                    rf.ps_flushlsn = buf.get_u64();
                }
                b"ps_applylsn" => {
                    let len = buf.get_i32();
                    assert_eq!(len, 8);
-                    rf.remote_consistent_lsn = buf.get_u64();
+                    rf.ps_applylsn = buf.get_u64();
                }
                b"ps_replytime" => {
                    let len = buf.get_i32();
                    assert_eq!(len, 8);
                    let raw_time = buf.get_i64();
                    if raw_time > 0 {
-                        rf.replytime = *PG_EPOCH + Duration::from_micros(raw_time as u64);
+                        rf.ps_replytime = *PG_EPOCH + Duration::from_micros(raw_time as u64);
                    } else {
-                        rf.replytime = *PG_EPOCH - Duration::from_micros(-raw_time as u64);
+                        rf.ps_replytime = *PG_EPOCH - Duration::from_micros(-raw_time as u64);
                    }
                }
                _ => {
                    let len = buf.get_i32();
                    warn!(
-                        "PageserverFeedback parse. unknown key {} of len {len}. Skip it.",
+                        "ReplicationFeedback parse. unknown key {} of len {len}. Skip it.",
                        String::from_utf8_lossy(key.as_ref())
                    );
                    buf.advance(len as usize);
                }
            }
        }
-        trace!("PageserverFeedback parsed is {:?}", rf);
+        trace!("ReplicationFeedback parsed is {:?}", rf);
        rf
    }
 }
@@ -1070,33 +1059,33 @@ mod tests {

    #[test]
    fn test_replication_feedback_serialization() {
-        let mut rf = PageserverFeedback::empty();
+        let mut rf = ReplicationFeedback::empty();
        // Fill rf with some values
        rf.current_timeline_size = 12345678;
        // Set rounded time to be able to compare it with deserialized value,
        // because it is rounded up to microseconds during serialization.
-        rf.replytime = *PG_EPOCH + Duration::from_secs(100_000_000);
+        rf.ps_replytime = *PG_EPOCH + Duration::from_secs(100_000_000);
        let mut data = BytesMut::new();
        rf.serialize(&mut data);

-        let rf_parsed = PageserverFeedback::parse(data.freeze());
+        let rf_parsed = ReplicationFeedback::parse(data.freeze());
        assert_eq!(rf, rf_parsed);
    }

    #[test]
    fn test_replication_feedback_unknown_key() {
-        let mut rf = PageserverFeedback::empty();
+        let mut rf = ReplicationFeedback::empty();
        // Fill rf with some values
        rf.current_timeline_size = 12345678;
        // Set rounded time to be able to compare it with deserialized value,
        // because it is rounded up to microseconds during serialization.
-        rf.replytime = *PG_EPOCH + Duration::from_secs(100_000_000);
+        rf.ps_replytime = *PG_EPOCH + Duration::from_secs(100_000_000);
        let mut data = BytesMut::new();
        rf.serialize(&mut data);

        // Add an extra field to the buffer and adjust number of keys
        if let Some(first) = data.first_mut() {
-            *first = PAGESERVER_FEEDBACK_FIELDS_NUMBER + 1;
+            *first = REPLICATION_FEEDBACK_FIELDS_NUMBER + 1;
        }

        data.put_slice(b"new_field_one\0");
@@ -1104,7 +1093,7 @@ mod tests {
        data.put_u64(42);

        // Parse serialized data and check that new field is not parsed
-        let rf_parsed = PageserverFeedback::parse(data.freeze());
+        let rf_parsed = ReplicationFeedback::parse(data.freeze());
        assert_eq!(rf, rf_parsed);
    }

--- a/libs/remote_storage/src/lib.rs
+++ b/libs/remote_storage/src/lib.rs
@@ -13,6 +13,7 @@ use std::{
    collections::HashMap,
    fmt::Debug,
    num::{NonZeroU32, NonZeroUsize},
+    ops::Deref,
    path::{Path, PathBuf},
    pin::Pin,
    sync::Arc,
@@ -77,6 +78,9 @@ impl RemotePath {
 /// providing basic CRUD operations for storage files.
 #[async_trait::async_trait]
 pub trait RemoteStorage: Send + Sync + 'static {
+    /// Lists all items the storage has right now.
+    async fn list(&self) -> anyhow::Result<Vec<RemotePath>>;
+
    /// Lists all top level subdirectories for a given prefix
    /// Note: here we assume that if the prefix is passed it was obtained via remote_object_id
    /// which already takes into account any kind of global prefix (prefix_in_bucket for S3 or storage_root for LocalFS)
@@ -89,7 +93,7 @@ pub trait RemoteStorage: Send + Sync + 'static {
    /// Streams the local file contents into remote into the remote storage entry.
    async fn upload(
        &self,
-        from: impl io::AsyncRead + Unpin + Send + Sync + 'static,
+        data: Box<(dyn io::AsyncRead + Unpin + Send + Sync + 'static)>,
        // S3 PUT request requires the content length to be specified,
        // otherwise it starts to fail with the concurrent connection count increasing.
        data_size_bytes: usize,
@@ -160,67 +164,14 @@ pub enum GenericRemoteStorage {
    Unreliable(Arc<UnreliableWrapper>),
 }

-impl GenericRemoteStorage {
-    pub async fn list_prefixes(
-        &self,
-        prefix: Option<&RemotePath>,
-    ) -> Result<Vec<RemotePath>, DownloadError> {
-        match self {
-            Self::LocalFs(s) => s.list_prefixes(prefix).await,
-            Self::AwsS3(s) => s.list_prefixes(prefix).await,
-            Self::Unreliable(s) => s.list_prefixes(prefix).await,
-        }
-    }
+impl Deref for GenericRemoteStorage {
+    type Target = dyn RemoteStorage;

-    pub async fn upload(
-        &self,
-        from: impl io::AsyncRead + Unpin + Send + Sync + 'static,
-        data_size_bytes: usize,
-        to: &RemotePath,
-        metadata: Option<StorageMetadata>,
-    ) -> anyhow::Result<()> {
+    fn deref(&self) -> &Self::Target {
        match self {
-            Self::LocalFs(s) => s.upload(from, data_size_bytes, to, metadata).await,
-            Self::AwsS3(s) => s.upload(from, data_size_bytes, to, metadata).await,
-            Self::Unreliable(s) => s.upload(from, data_size_bytes, to, metadata).await,
-        }
-    }
-
-    pub async fn download(&self, from: &RemotePath) -> Result<Download, DownloadError> {
-        match self {
-            Self::LocalFs(s) => s.download(from).await,
-            Self::AwsS3(s) => s.download(from).await,
-            Self::Unreliable(s) => s.download(from).await,
-        }
-    }
-
-    pub async fn download_byte_range(
-        &self,
-        from: &RemotePath,
-        start_inclusive: u64,
-        end_exclusive: Option<u64>,
-    ) -> Result<Download, DownloadError> {
-        match self {
-            Self::LocalFs(s) => {
-                s.download_byte_range(from, start_inclusive, end_exclusive)
-                    .await
-            }
-            Self::AwsS3(s) => {
-                s.download_byte_range(from, start_inclusive, end_exclusive)
-                    .await
-            }
-            Self::Unreliable(s) => {
-                s.download_byte_range(from, start_inclusive, end_exclusive)
-                    .await
-            }
-        }
-    }
-
-    pub async fn delete(&self, path: &RemotePath) -> anyhow::Result<()> {
-        match self {
-            Self::LocalFs(s) => s.delete(path).await,
-            Self::AwsS3(s) => s.delete(path).await,
-            Self::Unreliable(s) => s.delete(path).await,
+            GenericRemoteStorage::LocalFs(local_fs) => local_fs,
+            GenericRemoteStorage::AwsS3(s3_bucket) => s3_bucket.as_ref(),
+            GenericRemoteStorage::Unreliable(s) => s.as_ref(),
        }
    }
 }
@@ -251,7 +202,7 @@ impl GenericRemoteStorage {
    /// this path is used for the remote object id conversion only.
    pub async fn upload_storage_object(
        &self,
-        from: impl tokio::io::AsyncRead + Unpin + Send + Sync + 'static,
+        from: Box<dyn tokio::io::AsyncRead + Unpin + Send + Sync + 'static>,
        from_size_bytes: usize,
        to: &RemotePath,
    ) -> anyhow::Result<()> {
--- a/libs/remote_storage/src/local_fs.rs
+++ b/libs/remote_storage/src/local_fs.rs
@@ -73,8 +73,10 @@ impl LocalFs {
            Ok(None)
        }
    }
+}

-    #[cfg(test)]
+#[async_trait::async_trait]
+impl RemoteStorage for LocalFs {
    async fn list(&self) -> anyhow::Result<Vec<RemotePath>> {
        Ok(get_all_files(&self.storage_root, true)
            .await?
@@ -89,10 +91,7 @@ impl LocalFs {
            })
            .collect())
    }
-}

-#[async_trait::async_trait]
-impl RemoteStorage for LocalFs {
    async fn list_prefixes(
        &self,
        prefix: Option<&RemotePath>,
@@ -118,7 +117,7 @@ impl RemoteStorage for LocalFs {

    async fn upload(
        &self,
-        data: impl io::AsyncRead + Unpin + Send + Sync + 'static,
+        data: Box<(dyn io::AsyncRead + Unpin + Send + Sync + 'static)>,
        data_size_bytes: usize,
        to: &RemotePath,
        metadata: Option<StorageMetadata>,
--- a/libs/remote_storage/src/s3_bucket.rs
+++ b/libs/remote_storage/src/s3_bucket.rs
@@ -275,6 +275,50 @@ impl<S: AsyncRead> AsyncRead for RatelimitedAsyncRead<S> {

 #[async_trait::async_trait]
 impl RemoteStorage for S3Bucket {
+    async fn list(&self) -> anyhow::Result<Vec<RemotePath>> {
+        let mut document_keys = Vec::new();
+
+        let mut continuation_token = None;
+        loop {
+            let _guard = self
+                .concurrency_limiter
+                .acquire()
+                .await
+                .context("Concurrency limiter semaphore got closed during S3 list")?;
+
+            metrics::inc_list_objects();
+
+            let fetch_response = self
+                .client
+                .list_objects_v2()
+                .bucket(self.bucket_name.clone())
+                .set_prefix(self.prefix_in_bucket.clone())
+                .delimiter(REMOTE_STORAGE_PREFIX_SEPARATOR.to_string())
+                .set_continuation_token(continuation_token)
+                .set_max_keys(self.max_keys_per_list_response)
+                .send()
+                .await
+                .map_err(|e| {
+                    metrics::inc_list_objects_fail();
+                    e
+                })?;
+            document_keys.extend(
+                fetch_response
+                    .contents
+                    .unwrap_or_default()
+                    .into_iter()
+                    .filter_map(|o| Some(self.s3_object_to_relative_path(o.key()?))),
+            );
+
+            match fetch_response.next_continuation_token {
+                Some(new_token) => continuation_token = Some(new_token),
+                None => break,
+            }
+        }
+
+        Ok(document_keys)
+    }
+
    /// See the doc for `RemoteStorage::list_prefixes`
    /// Note: it wont include empty "directories"
    async fn list_prefixes(
@@ -343,7 +387,7 @@ impl RemoteStorage for S3Bucket {

    async fn upload(
        &self,
-        from: impl io::AsyncRead + Unpin + Send + Sync + 'static,
+        from: Box<(dyn io::AsyncRead + Unpin + Send + Sync + 'static)>,
        from_size_bytes: usize,
        to: &RemotePath,
        metadata: Option<StorageMetadata>,
--- a/libs/remote_storage/src/simulate_failures.rs
+++ b/libs/remote_storage/src/simulate_failures.rs
@@ -20,6 +20,7 @@ pub struct UnreliableWrapper {
 /// Used to identify retries of different unique operation.
 #[derive(Debug, Hash, Eq, PartialEq)]
 enum RemoteOp {
+    List,
    ListPrefixes(Option<RemotePath>),
    Upload(RemotePath),
    Download(RemotePath),
@@ -74,6 +75,12 @@ impl UnreliableWrapper {

 #[async_trait::async_trait]
 impl RemoteStorage for UnreliableWrapper {
+    /// Lists all items the storage has right now.
+    async fn list(&self) -> anyhow::Result<Vec<RemotePath>> {
+        self.attempt(RemoteOp::List)?;
+        self.inner.list().await
+    }
+
    async fn list_prefixes(
        &self,
        prefix: Option<&RemotePath>,
@@ -84,7 +91,7 @@ impl RemoteStorage for UnreliableWrapper {

    async fn upload(
        &self,
-        data: impl tokio::io::AsyncRead + Unpin + Send + Sync + 'static,
+        data: Box<(dyn tokio::io::AsyncRead + Unpin + Send + Sync + 'static)>,
        // S3 PUT request requires the content length to be specified,
        // otherwise it starts to fail with the concurrent connection count increasing.
        data_size_bytes: usize,
--- a/libs/timeline_data_path/Cargo.toml
+++ b/libs/timeline_data_path/Cargo.toml
@@ -1,13 +0,0 @@
-[package]
-name = "timeline_data_path"
-version = "0.1.0"
-edition.workspace = true
-license.workspace = true
-
-# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
-
-[dependencies]
-utils.workspace = true
-workspace_hack.workspace = true
-tokio.workspace = true
-thiserror.workspace = true
--- a/libs/timeline_data_path/src/lib.rs
+++ b/libs/timeline_data_path/src/lib.rs
@@ -1,396 +0,0 @@
-//! The Timeline's core data path.
-//!
-//! # Overview
-//!
-//! This crate implements the core data path of a Timeline inside Pageserver:
-//!
-//! 1. WAL records from `walreceiver`, via in-memory layers, into persistent L0 layers.
-//! 1. `GetPage@LSN`: retrieval of WAL records and page images for feeding into WAL redo.
-//! 1. Data re-shuffeling through compaction (TODO).
-//! 1. Page image creation & garbage collection through GC (TODO).
-//!
-//! This crate assumes the following concepts, but is fully generic over their implementation:
-//!
-//! - **Delta Records**: data is written into the system in the form of self-descriptive deltas.
-//!   For the Pageserver use case, these deltas are derived from Postgres WAL records.
-//! - **Page Numbers**: Delta Records always affect a single key.
-//!   That key is called page number, because, in the Pageserver use case, the Postgres table page numbers are the keys.
-//! - **LSN**: When writing Delta Records into the system, they are associated with a monotonically increasing LSN.
-//!   Subsequently written Delta Records must have increasing LSNs.
-//! - **Page Images**: Delta Records for a given page can be used to reconstruct the page. Think of it like squashing diffs.
-//!   - When sorting the Delta Records for a given key by their LSN, any prefix of that sorting can be squashed into a page image.
-//!   - Delta Records following such a squash can be squashed into that page image.
-//!   - In Pageserver, WAL redo implements the (pure) function of squashing.
-//! - **In-Memory Layer**: an object that represents an "unfinished" L0 layer file, holding Delta Records in insertion order.
-//!   "Unfinished" means that we're still writing Delta Records to that file.
-//! - **Historic Layer**: an object that represents a "finished" layer file, at any compaction level.
-//!   Such objects reside on disk and/or in remote storage.
-//!   They may contain Delta Records, Page Images, or a mixture thereof. It doesn't matter.
-//! - **HistoricStuff**: an efficient lookup data structure to find the list of Historic Layer objects
-//!   that hold the Delta Records / PageImages required to reconstruct a Page Image at a given LSN.
-//!
-//! # API
-//!
-//! The core idea is that of a specialized single-producer multi-consumer structure,
-//! embodied by a Read-end and a Write-end.
-//!
-//! The Write-end is used to push new `DeltaRecord @ LSN`s into the system.
-//! In Pageserver, this is used by the `WalReceiver`.
-//!
-//! The Read-end provides the `GetPage@LSN` API.
-//! In the current iteration, we actually return something called `ReconstructWork`.
-//! I.e., we leave the work of reading the values from the layers, and the WAL redo invocation to the caller.
-//! Find rationale for this design in the *Scope* section.
-//!
-//! ## Immutability
-//!
-//! The traits defined by this crate assume immutable data structures that are multi-versioned.
-//!
-//! As an example for what "immutable" means, take the case where we add a new Historic Layer to HistoricStuff.
-//! Traditionally, one would use shared mutable state, i.e. `Arc<RwLock<...>>`.
-//! To insert the new Historic Layer, we would acquire the RwLock in write mode and modify a lookup data structure to accomodate the new layer.
-//! The Read-ends would use RwLock in read mode to read from the data structure.
-//!
-//! Conversely, with  *immutable data structures*, writers create new version (aka *snapshots*) of the lookup data structure.
-//! New reads on the Read-ends will use the new snapshot, but old ongoing reads would use the old version(s).
-//! An efficient implementation would likely share the Historic Layer objects, e.g., using `Arc`.
-//! And maybe there's internally mutable state inside the layer objects, e.g., to track residence (i.e., *on-demand downloaded* vs *evicted*).
-//! But the important point is that there's no synchronization / lock-holding at any higher level, except when grabbing a reference to the snapshot (Read-end), or when publishing a new snapshot (Write-end).
-//!
-//! ## Scope
-//!
-//! The following concerns are considered implementation details from the perspective of this crate:
-//!
-//! - **Layer File Persistence**: `HistoricStuff::make_historic` is responsible for this.
-//! - **Reading Layer Files**: the `ReconstructWork` that the Read-end returns from `GetPage@LSN` requests contains the list of layers to consult.
-//!   The crate consumer is responsible for reading the layers & doing WAL redo.
-//!   Likely the implementation of `HistoricStuff` plays a role here, because it is responsible for persisting the layer files.
-//! - **Layer Eviction & On-Demand Download**: this is just an aspect of the above.
-//!   The crate consumer can choose to implement eviction & on-demand download however they wish.
-//!   The only requirement is that the Historic Layers don't change their contents, i.e., they always returnt he same reconstruct values for the same lookup.
-//!   - For example, a `LayerCache` modoule or service could take care of layer uploads, eviction, and on-demand downloads.
-//!     Initially, the `layer cache` can be local-only.
-//!     But in the future, it can be multi-machine / clustered pagesevers / aka "sharding".
-//!
-//! # Example
-//!
-//! The [`new`] function is the entrypoint to this crate.
-//!
-//! See the test cases for how it is used.
-
-use std::{marker::PhantomData, time::Duration};
-
-use utils::seqwait::{self, Advance, SeqWait, Wait};
-
-#[cfg(test)]
-mod tests;
-
-/// Collection of types / type bounds used by Read-end and Write-end.
-///
-/// See the [`crate`]-level docs's *Concepts* section to learn about
-/// the meaning of each associated `type`.
-///
-/// # Usage
-///
-/// Define a zero-sized-type and impl this Trait for it.
-/// Then use that zero-sized-type as the single generic argument to [`new`]
-/// and almost all types declared in this crate.
-///
-/// It might feel a bit weird, but, the alternative is to have umpteen generic
-/// types per `impl` with repetitive trait bounds.
-///
-/// Search the test cases for an example of how this can be used to improve testability.
-pub trait Types {
-    type Key: Copy;
-    type Lsn: Ord + Copy;
-    type LsnCounter: seqwait::MonotonicCounter<Self::Lsn> + Copy;
-    type DeltaRecord;
-    type HistoricLayer;
-    type InMemoryLayer: InMemoryLayer<Types = Self> + Clone;
-    type HistoricStuff: HistoricStuff<Types = Self> + Clone;
-    type GetReconstructPathError: std::error::Error;
-}
-
-/// Error returned by [`InMemoryLayer::put`].
-#[derive(thiserror::Error)]
-pub struct InMemoryLayerPutError<DeltaRecord> {
-    delta: DeltaRecord,
-    kind: InMemoryLayerPutErrorKind,
-}
-
-/// Part of [`InMemoryLayerPutError`].
-#[derive(Debug)]
-pub enum InMemoryLayerPutErrorKind {
-    LayerFull,
-    AlreadyHaveRecordForKeyAndLsn,
-}
-
-impl<DeltaRecord> std::fmt::Debug for InMemoryLayerPutError<DeltaRecord> {
-    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
-        f.debug_struct("InMemoryLayerPutError")
-            // would require DeltaRecord to impl Debug
-            //         .field("delta", &self.delta)
-            .field("kind", &self.kind)
-            .finish()
-    }
-}
-
-/// An in-memory layer. See [`crate`] docs for details on this concept.
-pub trait InMemoryLayer: std::fmt::Debug + Default + Clone {
-    type Types: Types;
-    fn put(
-        &mut self,
-        key: <Self::Types as Types>::Key,
-        lsn: <Self::Types as Types>::Lsn,
-        delta: <Self::Types as Types>::DeltaRecord,
-    ) -> Result<Self, InMemoryLayerPutError<<Self::Types as Types>::DeltaRecord>>;
-    fn get(
-        &self,
-        key: <Self::Types as Types>::Key,
-        lsn: <Self::Types as Types>::Lsn,
-    ) -> Vec<<Self::Types as Types>::DeltaRecord>;
-}
-
-/// The manager of [`Types::HistoricLayer`]s.
-pub trait HistoricStuff {
-    type Types: Types;
-    fn get_reconstruct_path(
-        &self,
-        key: <Self::Types as Types>::Key,
-        lsn: <Self::Types as Types>::Lsn,
-    ) -> Result<
-        Vec<<Self::Types as Types>::HistoricLayer>,
-        <Self::Types as Types>::GetReconstructPathError,
-    >;
-    /// Produce a new version of `self` that includes the given inmem layer.
-    fn make_historic(&self, inmem: <Self::Types as Types>::InMemoryLayer) -> Self;
-}
-
-/// A snapshot of the data. See [`crate`]-level docs section on *immutability* for details.
-struct Snapshot<T: Types> {
-    _types: PhantomData<T>,
-    inmem: Option<T::InMemoryLayer>,
-    historic: T::HistoricStuff,
-}
-
-impl<T: Types> Clone for Snapshot<T> {
-    fn clone(&self) -> Self {
-        Self {
-            _types: self._types.clone(),
-            inmem: self.inmem.clone(),
-            historic: self.historic.clone(),
-        }
-    }
-}
-
-/// The Read-end. See [`crate`]-level docs for details.
-pub struct Reader<T: Types> {
-    wait: Wait<T::LsnCounter, T::Lsn, Snapshot<T>>,
-}
-
-/// The Write-end. See [`crate`]-level docs for details.
-pub struct Writer<T: Types> {
-    advance: Advance<T::LsnCounter, T::Lsn, Snapshot<T>>,
-}
-
-/// Setup a pair of Read-end and Write-End. This is the entrypoint to this crate.
-///
-/// The idea is that the caller loads the arguments from persistent state that `HistoricStuff` wrote at an earlier point in time.
-pub fn new<T: Types>(lsn: T::LsnCounter, historic: T::HistoricStuff) -> (Reader<T>, Writer<T>) {
-    let state = Snapshot {
-        _types: PhantomData::<T>::default(),
-        inmem: None,
-        historic: historic,
-    };
-    let (wait, advance) = SeqWait::new(lsn, state).split_spmc();
-    let reader = Reader { wait };
-    let read_writer = Writer { advance };
-    (reader, read_writer)
-}
-
-/// Error returned by the get-page operations.
-#[derive(Debug, thiserror::Error)]
-pub enum GetError<T: Types> {
-    #[error(transparent)]
-    SeqWait(seqwait::SeqWaitError),
-    #[error(transparent)]
-    GetReconstructPath(T::GetReconstructPathError),
-}
-
-/// Self-contained set of objects required to reconstruct a page image for the given `key` @ `lsn`.
-///
-/// This is returned by the `get` methods of [`Reader`] and [`Writer`].
-///
-/// To reconstruct the page image, stack up (top to bottom) `inmem_records` plus all records found for `key` and `lsn` along the `historic_path` until an initial page image is found.
-/// Then feed that stack to WAL-redo to get the page image.
-///
-/// See [`crate`]-level docs on *scope* for why we don't return page images from these functions.
-pub struct ReconstructWork<T: Types> {
-    pub key: T::Key,
-    pub lsn: T::Lsn,
-    pub inmem_records: Vec<T::DeltaRecord>,
-    pub historic_path: Vec<T::HistoricLayer>,
-}
-
-impl<T: Types> Reader<T> {
-    /// This is the `GetPage@LSN` operation.
-    ///
-    /// See the [`crate`]-level docs for why we return [`ReconstructWork`] instead of a Page Image here.
-    pub async fn get(&self, key: T::Key, lsn: T::Lsn) -> Result<ReconstructWork<T>, GetError<T>> {
-        // XXX dedup with Writer::get_nowait
-        let state = self.wait.wait_for(lsn).await.map_err(GetError::SeqWait)?;
-        let inmem_records = state
-            .inmem
-            .as_ref()
-            .map(|iml| iml.get(key, lsn))
-            .unwrap_or_default();
-        let historic_path = state
-            .historic
-            .get_reconstruct_path(key, lsn)
-            .map_err(GetError::GetReconstructPath)?;
-        Ok(ReconstructWork {
-            key,
-            lsn,
-            inmem_records,
-            historic_path,
-        })
-    }
-}
-
-/// Error returned by the `put` operation.
-#[derive(thiserror::Error)]
-pub struct PutError<T: Types> {
-    /// The `delta` record which we failed to `put`.
-    pub delta: T::DeltaRecord,
-    /// Description of what went wrong.
-    pub kind: PutErrorKind,
-}
-
-/// Part of [`PutError`].
-#[derive(Debug)]
-pub enum PutErrorKind {
-    AlreadyHaveInMemoryRecordForKeyAndLsn,
-}
-
-impl<T: Types> std::fmt::Debug for PutError<T> {
-    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
-        f.debug_struct("PutError")
-            // would need to require Debug for DeltaRecord
-            // .field("delta", &self.delta)
-            .field("kind", &self.kind)
-            .finish()
-    }
-}
-
-impl<T: Types> Writer<T> {
-    /// Insert data into the system.
-    pub async fn put(
-        &mut self,
-        key: T::Key,
-        lsn: T::Lsn,
-        delta: T::DeltaRecord,
-    ) -> Result<(), PutError<T>> {
-        let (_snapshot_lsn, snapshot) = self.advance.get_current_data();
-        // TODO ensure snapshot_lsn <= lsn?
-        let mut inmem = snapshot
-            .inmem
-            .unwrap_or_else(|| T::InMemoryLayer::default());
-        // XXX: use the Advance as witness and only allow witness to access inmem in write mode
-        match inmem.put(key, lsn, delta) {
-            Ok(new_inmem) => {
-                let new_snapshot = Snapshot {
-                    _types: PhantomData,
-                    inmem: Some(new_inmem),
-                    historic: snapshot.historic,
-                };
-                self.advance.advance(lsn, Some(new_snapshot));
-            }
-            Err(InMemoryLayerPutError {
-                delta,
-                kind: InMemoryLayerPutErrorKind::AlreadyHaveRecordForKeyAndLsn,
-            }) => {
-                return Err(PutError {
-                    delta,
-                    kind: PutErrorKind::AlreadyHaveInMemoryRecordForKeyAndLsn,
-                });
-            }
-            Err(InMemoryLayerPutError {
-                delta,
-                kind: InMemoryLayerPutErrorKind::LayerFull,
-            }) => {
-                let new_historic = snapshot.historic.make_historic(inmem);
-                let mut new_inmem = T::InMemoryLayer::default();
-                let new_inmem = new_inmem
-                    .put(key, lsn, delta)
-                    .expect("put into default inmem layer must not fail");
-                let new_state = Snapshot {
-                    _types: PhantomData::<T>::default(),
-                    inmem: Some(new_inmem),
-                    historic: new_historic,
-                };
-                self.advance.advance(lsn, Some(new_state));
-            }
-        }
-        Ok(())
-    }
-
-    /// Force flushing of the current in-memory layer.
-    ///
-    /// Usually, flushing happens only if the in-memory layer is full.
-    /// Use this API to make it happen in other circumstances (shutdown, periodic ticker, etc.).
-    pub async fn force_flush(&mut self) -> tokio::io::Result<()> {
-        let (snapshot_lsn, snapshot) = self.advance.get_current_data();
-        let Snapshot {
-            _types,
-            inmem,
-            historic,
-        } = snapshot;
-        // XXX: use the Advance as witness and only allow witness to access inmem in "write" mode
-        let Some(inmem) = inmem else {
-            // nothing to do
-            return Ok(());
-        };
-        let new_historic = historic.make_historic(inmem);
-        let new_snapshot = Snapshot {
-            _types: PhantomData::<T>::default(),
-            inmem: None,
-            historic: new_historic,
-        };
-        self.advance.advance(snapshot_lsn, Some(new_snapshot)); // TODO: should fail if we're past snapshot_lsn
-        Ok(())
-    }
-
-    /// `get` at the given LSN, without blocking.
-    ///
-    /// Fails with a timeout error if the `lsn` isn't there yet.
-    /// That makes sense because the only way we'd stop waiting is by a `self.put()`.
-    /// But concurrent `put()` is forbidden.
-    pub async fn get_nowait(
-        &self,
-        key: T::Key,
-        lsn: T::Lsn,
-    ) -> Result<ReconstructWork<T>, GetError<T>> {
-        // XXX dedup with Reader::get
-        let state = self
-            .advance
-            .wait_for_timeout(lsn, Duration::from_secs(0))
-            // The await is never going to block because we pass from_secs(0).
-            .await
-            .map_err(GetError::SeqWait)?;
-        let inmem_records = state
-            .inmem
-            .as_ref()
-            .map(|iml| iml.get(key, lsn))
-            .unwrap_or_default();
-        let historic_path = state
-            .historic
-            .get_reconstruct_path(key, lsn)
-            .map_err(GetError::GetReconstructPath)?;
-        Ok(ReconstructWork {
-            key,
-            lsn,
-            inmem_records,
-            historic_path,
-        })
-    }
-}
--- a/libs/timeline_data_path/src/tests.rs
+++ b/libs/timeline_data_path/src/tests.rs
@@ -1,170 +0,0 @@
-use std::collections::{btree_map::Entry, BTreeMap};
-use std::sync::Arc;
-use utils::seqwait;
-
-/// The ZST for which we impl the `super::Types` type collection trait.
-struct TestTypes;
-
-impl super::Types for TestTypes {
-    type Key = usize;
-
-    type Lsn = usize;
-
-    type LsnCounter = UsizeCounter;
-
-    type DeltaRecord = &'static str;
-
-    type HistoricLayer = Arc<TestHistoricLayer>;
-
-    type InMemoryLayer = TestInMemoryLayer;
-
-    type HistoricStuff = TestHistoricStuff;
-}
-
-/// For testing, our in-memory layer is a simple hashmap.
-#[derive(Clone, Default, Debug)]
-struct TestInMemoryLayer {
-    by_key: BTreeMap<usize, BTreeMap<usize, &'static str>>,
-}
-
-/// For testing, our historic layers are just in-memory layer objects with `frozen==true`.
-struct TestHistoricLayer(TestInMemoryLayer);
-
-/// This is the data structure that impls the `HistoricStuff` trait.
-#[derive(Default, Clone)]
-struct TestHistoricStuff {
-    by_key: BTreeMap<usize, BTreeMap<usize, Arc<TestHistoricLayer>>>,
-}
-
-/// `seqwait::MonotonicCounter` impl
-#[derive(Copy, Clone)]
-pub struct UsizeCounter(usize);
-
-// Our testing impl of HistoricStuff references the frozen InMemoryLayer objects
-// from all the (key,lsn) entries that it covers.
-// This mimics the (much more efficient) search tree in the real impl.
-impl super::HistoricStuff for TestHistoricStuff {
-    type Types = TestTypes;
-    fn get_reconstruct_path(
-        &self,
-        key: usize,
-        lsn: usize,
-    ) -> Result<Vec<Arc<TestHistoricLayer>>, super::GetReconstructPathError> {
-        let Some(bk) = self.by_key.get(&key) else {
-                return Ok(vec![]);
-            };
-        Ok(bk.range(..=lsn).rev().map(|(_, l)| Arc::clone(l)).collect())
-    }
-
-    fn make_historic(&self, inmem: TestInMemoryLayer) -> Self {
-        // For the purposes of testing, just turn the inmemory layer historic through the type system
-        let historic = Arc::new(TestHistoricLayer(inmem));
-        // Deep-copy
-        let mut copy = self.by_key.clone();
-        // Add the references to `inmem` to the deep-copied struct
-        for (k, v) in historic.0.by_key.iter() {
-            for (lsn, _deltas) in v.into_iter() {
-                let by_key = copy.entry(*k).or_default();
-                let overwritten = by_key.insert(*lsn, historic.clone());
-                assert!(matches!(overwritten, None), "layers must not overlap");
-            }
-        }
-        Self { by_key: copy }
-    }
-}
-
-impl super::InMemoryLayer for TestInMemoryLayer {
-    type Types = TestTypes;
-
-    fn put(
-        &mut self,
-        key: usize,
-        lsn: usize,
-        delta: &'static str,
-    ) -> Result<Self, super::InMemoryLayerPutError<&'static str>> {
-        let mut clone = self.clone();
-        drop(self);
-        let by_key = clone.by_key.entry(key).or_default();
-        match by_key.entry(lsn) {
-            Entry::Occupied(_record) => {
-                return Err(super::InMemoryLayerPutError {
-                    delta,
-                    kind: super::InMemoryLayerPutErrorKind::AlreadyHaveRecordForKeyAndLsn,
-                });
-            }
-            Entry::Vacant(vacant) => vacant.insert(delta),
-        };
-        Ok(clone)
-    }
-
-    fn get(&self, key: usize, lsn: usize) -> Vec<&'static str> {
-        let by_key = match self.by_key.get(&key) {
-            Some(by_key) => by_key,
-            None => return vec![],
-        };
-        by_key
-            .range(..=lsn)
-            .map(|(_, v)| v)
-            .rev()
-            .cloned()
-            .collect()
-    }
-}
-
-impl UsizeCounter {
-    pub fn new(inital: usize) -> Self {
-        UsizeCounter(inital)
-    }
-}
-
-impl seqwait::MonotonicCounter<usize> for UsizeCounter {
-    fn cnt_advance(&mut self, new_val: usize) {
-        assert!(self.0 < new_val);
-        self.0 = new_val;
-    }
-
-    fn cnt_value(&self) -> usize {
-        self.0
-    }
-}
-
-#[test]
-fn basic() {
-    let lm = TestHistoricStuff::default();
-
-    let (r, mut rw) = super::new::<TestTypes>(UsizeCounter::new(0), lm);
-
-    let r = Arc::new(r);
-    let r2 = Arc::clone(&r);
-
-    let rt = tokio::runtime::Builder::new_current_thread()
-        .enable_all()
-        .build()
-        .unwrap();
-
-    let read_jh = rt.spawn(async move { r.get(0, 10).await });
-
-    let mut rw = rt.block_on(async move {
-        rw.put(0, 1, "foo").await.unwrap();
-        rw.put(1, 1, "bar").await.unwrap();
-        rw.put(0, 10, "baz").await.unwrap();
-        rw
-    });
-
-    let read_res = rt.block_on(read_jh).unwrap().unwrap();
-    assert!(
-        read_res.historic_path.is_empty(),
-        "we have pushed less than needed for flush"
-    );
-    assert_eq!(read_res.inmem_records, vec!["baz", "foo"]);
-
-    let rw = rt.block_on(async move {
-        rw.put(0, 11, "blup").await.unwrap();
-        rw
-    });
-    let read_res = rt.block_on(async move { r2.get(0, 11).await.unwrap() });
-    assert_eq!(read_res.historic_path.len(), 0);
-    assert_eq!(read_res.inmem_records, vec!["blup", "baz", "foo"]);
-
-    drop(rw);
-}
--- a/libs/utils/Cargo.toml
+++ b/libs/utils/Cargo.toml
@@ -19,7 +19,6 @@ jsonwebtoken.workspace = true
 nix.workspace = true
 once_cell.workspace = true
 pin-project-lite.workspace = true
-regex.workspace = true
 routerify.workspace = true
 serde.workspace = true
 serde_json.workspace = true
@@ -37,7 +36,6 @@ uuid = { version = "1.2", features = ["v4", "serde"] }

 metrics.workspace = true
 workspace_hack.workspace = true
-either.workspace = true

 [dev-dependencies]
 byteorder.workspace = true
--- a/libs/utils/src/lib.rs
+++ b/libs/utils/src/lib.rs
@@ -51,9 +51,6 @@ pub mod history_buffer;

 pub mod measured_stream;

-pub mod serde_percent;
-pub mod serde_regex;
-
 /// use with fail::cfg("$name", "return(2000)")
 #[macro_export]
 macro_rules! failpoint_sleep_millis_async {
--- a/libs/utils/src/seqwait.rs
+++ b/libs/utils/src/seqwait.rs
@@ -1,13 +1,12 @@
 #![warn(missing_docs)]

-use either::Either;
 use std::cmp::{Eq, Ordering, PartialOrd};
 use std::collections::BinaryHeap;
 use std::fmt::Debug;
 use std::mem;
-use std::sync::{Arc, Mutex};
+use std::sync::Mutex;
 use std::time::Duration;
-use tokio::sync::oneshot::{channel, Receiver, Sender};
+use tokio::sync::watch::{channel, Receiver, Sender};
 use tokio::time::timeout;

 /// An error happened while waiting for a number
@@ -37,48 +36,45 @@ pub trait MonotonicCounter<V> {
 }

 /// Internal components of a `SeqWait`
-struct SeqWaitInt<S, V, T>
+struct SeqWaitInt<S, V>
 where
    S: MonotonicCounter<V>,
    V: Ord,
-    T: Clone,
 {
-    waiters: BinaryHeap<Waiter<V, T>>,
+    waiters: BinaryHeap<Waiter<V>>,
    current: S,
    shutdown: bool,
-    data: T,
 }

-struct Waiter<V, T>
+struct Waiter<T>
 where
-    V: Ord,
-    T: Clone,
+    T: Ord,
 {
-    wake_num: V,             // wake me when this number arrives ...
-    wake_channel: Sender<T>, // ... by sending a message to this channel
+    wake_num: T,              // wake me when this number arrives ...
+    wake_channel: Sender<()>, // ... by sending a message to this channel
 }

 // BinaryHeap is a max-heap, and we want a min-heap. Reverse the ordering here
 // to get that.
-impl<V: Ord, T: Clone> PartialOrd for Waiter<V, T> {
+impl<T: Ord> PartialOrd for Waiter<T> {
    fn partial_cmp(&self, other: &Self) -> Option<Ordering> {
        other.wake_num.partial_cmp(&self.wake_num)
    }
 }

-impl<V: Ord, T: Clone> Ord for Waiter<V, T> {
+impl<T: Ord> Ord for Waiter<T> {
    fn cmp(&self, other: &Self) -> Ordering {
        other.wake_num.cmp(&self.wake_num)
    }
 }

-impl<V: Ord, T: Clone> PartialEq for Waiter<V, T> {
+impl<T: Ord> PartialEq for Waiter<T> {
    fn eq(&self, other: &Self) -> bool {
        other.wake_num == self.wake_num
    }
 }

-impl<V: Ord, T: Clone> Eq for Waiter<V, T> {}
+impl<T: Ord> Eq for Waiter<T> {}

 /// A tool for waiting on a sequence number
 ///
@@ -96,28 +92,25 @@ impl<V: Ord, T: Clone> Eq for Waiter<V, T> {}
 ///
 /// <S> means Storage, <V> is type of counter that this storage exposes.
 ///
-pub struct SeqWait<S, V, T>
+pub struct SeqWait<S, V>
 where
    S: MonotonicCounter<V>,
    V: Ord,
-    T: Clone,
 {
-    internal: Mutex<SeqWaitInt<S, V, T>>,
+    internal: Mutex<SeqWaitInt<S, V>>,
 }

-impl<S, V, T> SeqWait<S, V, T>
+impl<S, V> SeqWait<S, V>
 where
    S: MonotonicCounter<V> + Copy,
    V: Ord + Copy,
-    T: Clone,
 {
    /// Create a new `SeqWait`, initialized to a particular number
-    pub fn new(starting_num: S, data: T) -> Self {
+    pub fn new(starting_num: S) -> Self {
        let internal = SeqWaitInt {
            waiters: BinaryHeap::new(),
            current: starting_num,
            shutdown: false,
-            data,
        };
        SeqWait {
            internal: Mutex::new(internal),
@@ -151,13 +144,10 @@ where
    ///
    /// This call won't complete until someone has called `advance`
    /// with a number greater than or equal to the one we're waiting for.
-    pub async fn wait_for(&self, num: V) -> Result<T, SeqWaitError> {
-        match self.queue_for_wait(num, false) {
-            Ok(Either::Left(data)) => Ok(data),
-            Ok(Either::Right(rx)) => match rx.await {
-                Err(_) => Err(SeqWaitError::Shutdown),
-                Ok(data) => Ok(data),
-            },
+    pub async fn wait_for(&self, num: V) -> Result<(), SeqWaitError> {
+        match self.queue_for_wait(num) {
+            Ok(None) => Ok(()),
+            Ok(Some(mut rx)) => rx.changed().await.map_err(|_| SeqWaitError::Shutdown),
            Err(e) => Err(e),
        }
    }
@@ -169,18 +159,15 @@ where
    ///
    /// If that hasn't happened after the specified timeout duration,
    /// [`SeqWaitError::Timeout`] will be returned.
-    ///
-    /// Pass `timeout_duration.is_zero() == true` to guarantee that the
-    /// future that is this function will never await.
    pub async fn wait_for_timeout(
        &self,
        num: V,
        timeout_duration: Duration,
-    ) -> Result<T, SeqWaitError> {
-        match self.queue_for_wait(num, timeout_duration.is_zero()) {
-            Ok(Either::Left(data)) => Ok(data),
-            Ok(Either::Right(rx)) => match timeout(timeout_duration, rx).await {
-                Ok(Ok(data)) => Ok(data),
+    ) -> Result<(), SeqWaitError> {
+        match self.queue_for_wait(num) {
+            Ok(None) => Ok(()),
+            Ok(Some(mut rx)) => match timeout(timeout_duration, rx.changed()).await {
+                Ok(Ok(())) => Ok(()),
                Ok(Err(_)) => Err(SeqWaitError::Shutdown),
                Err(_) => Err(SeqWaitError::Timeout),
            },
@@ -190,50 +177,41 @@ where

    /// Register and return a channel that will be notified when a number arrives,
    /// or None, if it has already arrived.
-    fn queue_for_wait(&self, num: V, nowait: bool) -> Result<Either<T, Receiver<T>>, SeqWaitError> {
+    fn queue_for_wait(&self, num: V) -> Result<Option<Receiver<()>>, SeqWaitError> {
        let mut internal = self.internal.lock().unwrap();
        if internal.current.cnt_value() >= num {
-            return Ok(Either::Left(internal.data.clone()));
+            return Ok(None);
        }
        if internal.shutdown {
            return Err(SeqWaitError::Shutdown);
        }
-        if nowait {
-            return Err(SeqWaitError::Timeout);
-        }

        // Create a new channel.
-        let (tx, rx) = channel();
+        let (tx, rx) = channel(());
        internal.waiters.push(Waiter {
            wake_num: num,
            wake_channel: tx,
        });
        // Drop the lock as we exit this scope.
-        Ok(Either::Right(rx))
+        Ok(Some(rx))
    }

    /// Announce a new number has arrived
    ///
    /// All waiters at this value or below will be woken.
    ///
-    /// If `new_data` is Some(), it will update the internal data,
-    /// even if `num` is smaller than the internal counter.
-    /// It will not cause a wake-up though, in this case.
-    ///
    /// Returns the old number.
-    pub fn advance(&self, num: V, new_data: Option<T>) -> V {
+    pub fn advance(&self, num: V) -> V {
        let old_value;
-        let (wake_these, with_data) = {
+        let wake_these = {
            let mut internal = self.internal.lock().unwrap();
-            if let Some(new_data) = new_data {
-                internal.data = new_data;
-            }

            old_value = internal.current.cnt_value();
            if old_value >= num {
                return old_value;
            }
            internal.current.cnt_advance(num);
+
            // Pop all waiters <= num from the heap. Collect them in a vector, and
            // wake them up after releasing the lock.
            let mut wake_these = Vec::new();
@@ -243,13 +221,13 @@ where
                }
                wake_these.push(internal.waiters.pop().unwrap().wake_channel);
            }
-            (wake_these, internal.data.clone())
+            wake_these
        };

        for tx in wake_these {
            // This can fail if there are no receivers.
            // We don't care; discard the error.
-            let _ = tx.send(with_data.clone());
+            let _ = tx.send(());
        }
        old_value
    }
@@ -258,106 +236,6 @@ where
    pub fn load(&self) -> S {
        self.internal.lock().unwrap().current
    }
-
-    /// Split the seqwait into a part than can only do wait,
-    /// and another part that can do advance + wait.
-    ///
-    /// The wait-only part can be cloned, the advance part cannot be cloned.
-    /// This provides a single-producer multi-consumer scheme.
-    pub fn split_spmc(self) -> (Wait<S, V, T>, Advance<S, V, T>) {
-        let inner = Arc::new(self);
-        let w = Wait {
-            inner: inner.clone(),
-        };
-        let a = Advance { inner };
-        (w, a)
-    }
-}
-
-/// See [`SeqWait::split_spmc`].
-pub struct Wait<S, V, T>
-where
-    S: MonotonicCounter<V> + Copy,
-    V: Ord + Copy,
-    T: Clone,
-{
-    inner: Arc<SeqWait<S, V, T>>,
-}
-
-/// See [`SeqWait::split_spmc`].
-pub struct Advance<S, V, T>
-where
-    S: MonotonicCounter<V> + Copy,
-    V: Ord + Copy,
-    T: Clone,
-{
-    inner: Arc<SeqWait<S, V, T>>,
-}
-
-impl<S, V, T> Wait<S, V, T>
-where
-    S: MonotonicCounter<V> + Copy,
-    V: Ord + Copy,
-    T: Clone,
-{
-    /// See [`SeqWait::wait_for`].
-    pub async fn wait_for(&self, num: V) -> Result<T, SeqWaitError> {
-        self.inner.wait_for(num).await
-    }
-
-    /// See [`SeqWait::wait_for_timeout`].
-    pub async fn wait_for_timeout(
-        &self,
-        num: V,
-        timeout_duration: Duration,
-    ) -> Result<T, SeqWaitError> {
-        self.inner.wait_for_timeout(num, timeout_duration).await
-    }
-}
-
-impl<S, V, T> Advance<S, V, T>
-where
-    S: MonotonicCounter<V> + Copy,
-    V: Ord + Copy,
-    T: Clone,
-{
-    /// See [`SeqWait::advance`].
-    pub fn advance(&self, num: V, new_data: Option<T>) -> V {
-        self.inner.advance(num, new_data)
-    }
-
-    /// See [`SeqWait::wait_for`].
-    pub async fn wait_for(&self, num: V) -> Result<T, SeqWaitError> {
-        self.inner.wait_for(num).await
-    }
-
-    /// See [`SeqWait::wait_for_timeout`].
-    pub async fn wait_for_timeout(
-        &self,
-        num: V,
-        timeout_duration: Duration,
-    ) -> Result<T, SeqWaitError> {
-        self.inner.wait_for_timeout(num, timeout_duration).await
-    }
-
-    /// Get a `Clone::clone` of the current data inside the seqwait.
-    pub fn get_current_data(&self) -> (V, T) {
-        let inner = self.inner.internal.lock().unwrap();
-        (inner.current.cnt_value(), inner.data.clone())
-    }
-}
-
-impl<S, V, T> Clone for Wait<S, V, T>
-where
-    S: MonotonicCounter<V> + Copy,
-    V: Ord + Copy,
-    T: Clone,
-{
-    fn clone(&self) -> Self {
-        Self {
-            inner: self.inner.clone(),
-        }
-    }
 }

 #[cfg(test)]
@@ -378,12 +256,12 @@ mod tests {

    #[tokio::test]
    async fn seqwait() {
-        let seq = Arc::new(SeqWait::new(0, ()));
+        let seq = Arc::new(SeqWait::new(0));
        let seq2 = Arc::clone(&seq);
        let seq3 = Arc::clone(&seq);
        let jh1 = tokio::task::spawn(async move {
            seq2.wait_for(42).await.expect("wait_for 42");
-            let old = seq2.advance(100, None);
+            let old = seq2.advance(100);
            assert_eq!(old, 99);
            seq2.wait_for_timeout(999, Duration::from_millis(100))
                .await
@@ -394,12 +272,12 @@ mod tests {
            seq3.wait_for(0).await.expect("wait_for 0");
        });
        tokio::time::sleep(Duration::from_millis(200)).await;
-        let old = seq.advance(99, None);
+        let old = seq.advance(99);
        assert_eq!(old, 0);
        seq.wait_for(100).await.expect("wait_for 100");

        // Calling advance with a smaller value is a no-op
-        assert_eq!(seq.advance(98, None), 100);
+        assert_eq!(seq.advance(98), 100);
        assert_eq!(seq.load(), 100);

        jh1.await.unwrap();
@@ -410,7 +288,7 @@ mod tests {

    #[tokio::test]
    async fn seqwait_timeout() {
-        let seq = Arc::new(SeqWait::new(0, ()));
+        let seq = Arc::new(SeqWait::new(0));
        let seq2 = Arc::clone(&seq);
        let jh = tokio::task::spawn(async move {
            let timeout = Duration::from_millis(1);
@@ -420,104 +298,10 @@ mod tests {
        tokio::time::sleep(Duration::from_millis(200)).await;
        // This will attempt to wake, but nothing will happen
        // because the waiter already dropped its Receiver.
-        let old = seq.advance(99, None);
+        let old = seq.advance(99);
        assert_eq!(old, 0);
        jh.await.unwrap();

        seq.shutdown();
    }
-
-    #[tokio::test]
-    async fn data_basic() {
-        let seq = Arc::new(SeqWait::new(0, "a"));
-        let seq2 = Arc::clone(&seq);
-        let jh = tokio::task::spawn(async move {
-            let data = seq.wait_for(2).await.unwrap();
-            assert_eq!(data, "b");
-        });
-        seq2.advance(1, Some("x"));
-        seq2.advance(2, Some("b"));
-        jh.await.unwrap();
-    }
-
-    #[test]
-    fn data_always_most_recent() {
-        let rt = tokio::runtime::Builder::new_current_thread()
-            .build()
-            .unwrap();
-
-        let seq = Arc::new(SeqWait::new(0, "a"));
-        let seq2 = Arc::clone(&seq);
-
-        let jh = rt.spawn(async move {
-            let data = seq.wait_for(2).await.unwrap();
-            assert_eq!(data, "d");
-        });
-
-        // jh is not running until we poll it, thanks to current thread runtime
-
-        rt.block_on(async move {
-            seq2.advance(2, Some("b"));
-            seq2.advance(3, Some("c"));
-            seq2.advance(4, Some("d"));
-        });
-
-        rt.block_on(jh).unwrap();
-    }
-
-    #[tokio::test]
-    async fn split_spmc_api_surface() {
-        let seq = SeqWait::new(0, 1);
-        let (w, a) = seq.split_spmc();
-
-        let _ = w.wait_for(1);
-        let _ = w.wait_for_timeout(0, Duration::from_secs(10));
-        let _ = w.clone();
-
-        let _ = a.advance(1, None);
-        let _ = a.wait_for(1);
-        let _ = a.wait_for_timeout(0, Duration::from_secs(10));
-
-        // TODO would be nice to have must-not-compile tests for Advance not being clonable.
-    }
-
-    #[tokio::test]
-    async fn new_data_same_lsn() {
-        let seq = Arc::new(SeqWait::new(0, "a"));
-
-        seq.advance(1, Some("b"));
-        let data = seq.wait_for(1).await.unwrap();
-        assert_eq!(data, "b", "the regular case where lsn and data advance");
-
-        seq.advance(1, Some("c"));
-        let data = seq.wait_for(1).await.unwrap();
-        assert_eq!(
-            data, "c",
-            "no lsn advance still gives new data for old lsn wait_for's"
-        );
-
-        let (start_wait_for_sender, start_wait_for_receiver) = tokio::sync::oneshot::channel();
-        // ensure we don't wake waiters for data-only change
-        let jh = tokio::spawn({
-            let seq = seq.clone();
-            async move {
-                start_wait_for_receiver.await.unwrap();
-                match tokio::time::timeout(Duration::from_secs(2), seq.wait_for(2)).await {
-                    Ok(_) => {
-                        assert!(
-                            false,
-                            "advance should not wake waiters if data changes but LSN doesn't"
-                        );
-                    }
-                    Err(_) => {
-                        // Good, we weren't woken up.
-                    }
-                }
-            }
-        });
-
-        seq.advance(1, Some("d"));
-        start_wait_for_sender.send(()).unwrap();
-        jh.await.unwrap();
-    }
 }
--- a/libs/utils/src/serde_percent.rs
+++ b/libs/utils/src/serde_percent.rs
@@ -1,91 +0,0 @@
-//! A serde::Deserialize type for percentages.
-//!
-//! See [`Percent`] for details.
-
-use serde::{Deserialize, Serialize};
-
-/// If the value is not an integer between 0 and 100,
-/// deserialization fails with a descriptive error.
-#[derive(Debug, Clone, Copy, PartialEq, Eq, PartialOrd, Ord, Hash, Serialize, Deserialize)]
-#[serde(transparent)]
-pub struct Percent(#[serde(deserialize_with = "deserialize_pct_0_to_100")] u8);
-
-impl Percent {
-    pub const fn new(pct: u8) -> Option<Self> {
-        if pct <= 100 {
-            Some(Percent(pct))
-        } else {
-            None
-        }
-    }
-
-    pub fn get(&self) -> u8 {
-        self.0
-    }
-}
-
-fn deserialize_pct_0_to_100<'de, D>(deserializer: D) -> Result<u8, D::Error>
-where
-    D: serde::de::Deserializer<'de>,
-{
-    let v: u8 = serde::de::Deserialize::deserialize(deserializer)?;
-    if v > 100 {
-        return Err(serde::de::Error::custom(
-            "must be an integer between 0 and 100",
-        ));
-    }
-    Ok(v)
-}
-
-#[cfg(test)]
-mod tests {
-    use super::Percent;
-
-    #[derive(serde::Deserialize, serde::Serialize, Debug, PartialEq, Eq)]
-    struct Foo {
-        bar: Percent,
-    }
-
-    #[test]
-    fn basics() {
-        let input = r#"{ "bar": 50 }"#;
-        let foo: Foo = serde_json::from_str(input).unwrap();
-        assert_eq!(foo.bar.get(), 50);
-    }
-    #[test]
-    fn null_handling() {
-        let input = r#"{ "bar": null }"#;
-        let res: Result<Foo, _> = serde_json::from_str(input);
-        assert!(res.is_err());
-    }
-    #[test]
-    fn zero() {
-        let input = r#"{ "bar": 0 }"#;
-        let foo: Foo = serde_json::from_str(input).unwrap();
-        assert_eq!(foo.bar.get(), 0);
-    }
-    #[test]
-    fn out_of_range_above() {
-        let input = r#"{ "bar": 101 }"#;
-        let res: Result<Foo, _> = serde_json::from_str(input);
-        assert!(res.is_err());
-    }
-    #[test]
-    fn out_of_range_below() {
-        let input = r#"{ "bar": -1 }"#;
-        let res: Result<Foo, _> = serde_json::from_str(input);
-        assert!(res.is_err());
-    }
-    #[test]
-    fn float() {
-        let input = r#"{ "bar": 50.5 }"#;
-        let res: Result<Foo, _> = serde_json::from_str(input);
-        assert!(res.is_err());
-    }
-    #[test]
-    fn string() {
-        let input = r#"{ "bar": "50 %" }"#;
-        let res: Result<Foo, _> = serde_json::from_str(input);
-        assert!(res.is_err());
-    }
-}
--- a/libs/utils/src/serde_regex.rs
+++ b/libs/utils/src/serde_regex.rs
@@ -1,60 +0,0 @@
-//! A `serde::{Deserialize,Serialize}` type for regexes.
-
-use std::ops::Deref;
-
-#[derive(Debug, Clone, serde::Serialize, serde::Deserialize)]
-#[serde(transparent)]
-pub struct Regex(
-    #[serde(
-        deserialize_with = "deserialize_regex",
-        serialize_with = "serialize_regex"
-    )]
-    regex::Regex,
-);
-
-fn deserialize_regex<'de, D>(deserializer: D) -> Result<regex::Regex, D::Error>
-where
-    D: serde::de::Deserializer<'de>,
-{
-    let s: String = serde::de::Deserialize::deserialize(deserializer)?;
-    let re = regex::Regex::new(&s).map_err(serde::de::Error::custom)?;
-    Ok(re)
-}
-
-fn serialize_regex<S>(re: &regex::Regex, serializer: S) -> Result<S::Ok, S::Error>
-where
-    S: serde::ser::Serializer,
-{
-    serializer.collect_str(re.as_str())
-}
-
-impl Deref for Regex {
-    type Target = regex::Regex;
-
-    fn deref(&self) -> &regex::Regex {
-        &self.0
-    }
-}
-
-impl PartialEq for Regex {
-    fn eq(&self, other: &Regex) -> bool {
-        // comparing the automatons would be quite complicated
-        self.as_str() == other.as_str()
-    }
-}
-
-impl Eq for Regex {}
-
-#[cfg(test)]
-mod tests {
-
-    #[test]
-    fn roundtrip() {
-        let input = r#""foo.*bar""#;
-        let re: super::Regex = serde_json::from_str(input).unwrap();
-        assert!(re.is_match("foo123bar"));
-        assert!(!re.is_match("foo"));
-        let output = serde_json::to_string(&re).unwrap();
-        assert_eq!(output, input);
-    }
-}
--- a/pageserver/Cargo.toml
+++ b/pageserver/Cargo.toml
@@ -48,7 +48,6 @@ serde_json = { workspace = true, features = ["raw_value"] }
 serde_with.workspace = true
 signal-hook.workspace = true
 svg_fmt.workspace = true
-sync_wrapper.workspace = true
 tokio-tar.workspace = true
 thiserror.workspace = true
 tokio = { workspace = true, features = ["process", "sync", "fs", "rt", "io-util", "time"] }
--- a/pageserver/benches/bench_layer_map.rs
+++ b/pageserver/benches/bench_layer_map.rs
@@ -33,7 +33,7 @@ fn build_layer_map(filename_dump: PathBuf) -> LayerMap<LayerDescriptor> {
        min_lsn = min(min_lsn, lsn_range.start);
        max_lsn = max(max_lsn, Lsn(lsn_range.end.0 - 1));

-        updates.insert_historic(Arc::new(layer)).unwrap();
+        updates.insert_historic(Arc::new(layer));
    }

    println!("min: {min_lsn}, max: {max_lsn}");
@@ -215,7 +215,7 @@ fn bench_sequential(c: &mut Criterion) {
            is_incremental: false,
            short_id: format!("Layer {}", i),
        };
-        updates.insert_historic(Arc::new(layer)).unwrap();
+        updates.insert_historic(Arc::new(layer));
    }
    updates.flush();
    println!("Finished layer map init in {:?}", now.elapsed());
--- a/pageserver/src/bin/pageserver.rs
+++ b/pageserver/src/bin/pageserver.rs
@@ -8,7 +8,6 @@ use anyhow::{anyhow, Context};
 use clap::{Arg, ArgAction, Command};
 use fail::FailScenario;
 use metrics::launch_timestamp::{set_launch_timestamp_metric, LaunchTimestamp};
-use pageserver::disk_usage_eviction_task::{self, launch_disk_usage_global_eviction_task};
 use remote_storage::GenericRemoteStorage;
 use tracing::*;

@@ -315,34 +314,14 @@ fn start_pageserver(
    // Scan the local 'tenants/' directory and start loading the tenants
    BACKGROUND_RUNTIME.block_on(mgr::init_tenant_mgr(conf, remote_storage.clone()))?;

-    // shared state between the disk-usage backed eviction background task and the http endpoint
-    // that allows triggering disk-usage based eviction manually. note that the http endpoint
-    // is still accessible even if background task is not configured as long as remote storage has
-    // been configured.
-    let disk_usage_eviction_state: Arc<disk_usage_eviction_task::State> = Arc::default();
-
-    if let Some(remote_storage) = &remote_storage {
-        launch_disk_usage_global_eviction_task(
-            conf,
-            remote_storage.clone(),
-            disk_usage_eviction_state.clone(),
-        )?;
-    }
-
    // Start up the service to handle HTTP mgmt API request. We created the
    // listener earlier already.
    {
        let _rt_guard = MGMT_REQUEST_RUNTIME.enter();

-        let router = http::make_router(
-            conf,
-            launch_ts,
-            http_auth,
-            remote_storage,
-            disk_usage_eviction_state,
-        )?
-        .build()
-        .map_err(|err| anyhow!(err))?;
+        let router = http::make_router(conf, launch_ts, http_auth, remote_storage)?
+            .build()
+            .map_err(|err| anyhow!(err))?;
        let service = utils::http::RouterService::new(router).unwrap();
        let server = hyper::Server::from_tcp(http_listener)?
            .serve(service)
--- a/pageserver/src/config.rs
+++ b/pageserver/src/config.rs
@@ -27,7 +27,6 @@ use utils::{
    logging::LogFormat,
 };

-use crate::disk_usage_eviction_task::DiskUsageEvictionTaskConfig;
 use crate::tenant::config::TenantConf;
 use crate::tenant::config::TenantConfOpt;
 use crate::tenant::{TENANT_ATTACHING_MARKER_FILENAME, TIMELINES_SEGMENT_NAME};
@@ -93,8 +92,6 @@ pub mod defaults {

 #evictions_low_residence_duration_metric_threshold = '{DEFAULT_EVICTIONS_LOW_RESIDENCE_DURATION_METRIC_THRESHOLD}'

-#disk_usage_based_eviction = {{ max_usage_pct = .., min_avail_bytes = .., period = "10s"}}
-
 # [tenant_config]
 #checkpoint_distance = {DEFAULT_CHECKPOINT_DISTANCE} # in bytes
 #checkpoint_timeout = {DEFAULT_CHECKPOINT_TIMEOUT}
@@ -107,8 +104,6 @@ pub mod defaults {
 #image_creation_threshold = {DEFAULT_IMAGE_CREATION_THRESHOLD}
 #pitr_interval = '{DEFAULT_PITR_INTERVAL}'

-#min_resident_size_override = .. # in bytes
-
 # [remote_storage]

 "###
@@ -185,8 +180,6 @@ pub struct PageServerConf {
    // See the corresponding metric's help string.
    pub evictions_low_residence_duration_metric_threshold: Duration,

-    pub disk_usage_based_eviction: Option<DiskUsageEvictionTaskConfig>,
-
    pub test_remote_failures: u64,

    pub ondemand_download_behavior_treat_error_as_warn: bool,
@@ -259,8 +252,6 @@ struct PageServerConfigBuilder {

    evictions_low_residence_duration_metric_threshold: BuilderValue<Duration>,

-    disk_usage_based_eviction: BuilderValue<Option<DiskUsageEvictionTaskConfig>>,
-
    test_remote_failures: BuilderValue<u64>,

    ondemand_download_behavior_treat_error_as_warn: BuilderValue<bool>,
@@ -321,8 +312,6 @@ impl Default for PageServerConfigBuilder {
            )
            .expect("cannot parse DEFAULT_EVICTIONS_LOW_RESIDENCE_DURATION_METRIC_THRESHOLD")),

-            disk_usage_based_eviction: Set(None),
-
            test_remote_failures: Set(0),

            ondemand_download_behavior_treat_error_as_warn: Set(false),
@@ -442,10 +431,6 @@ impl PageServerConfigBuilder {
        self.evictions_low_residence_duration_metric_threshold = BuilderValue::Set(value);
    }

-    pub fn disk_usage_based_eviction(&mut self, value: Option<DiskUsageEvictionTaskConfig>) {
-        self.disk_usage_based_eviction = BuilderValue::Set(value);
-    }
-
    pub fn ondemand_download_behavior_treat_error_as_warn(
        &mut self,
        ondemand_download_behavior_treat_error_as_warn: bool,
@@ -530,9 +515,6 @@ impl PageServerConfigBuilder {
                .ok_or(anyhow!(
                    "missing evictions_low_residence_duration_metric_threshold"
                ))?,
-            disk_usage_based_eviction: self
-                .disk_usage_based_eviction
-                .ok_or(anyhow!("missing disk_usage_based_eviction"))?,
            test_remote_failures: self
                .test_remote_failures
                .ok_or(anyhow!("missing test_remote_failuers"))?,
@@ -722,12 +704,6 @@ impl PageServerConf {
                    builder.synthetic_size_calculation_interval(parse_toml_duration(key, item)?),
                "test_remote_failures" => builder.test_remote_failures(parse_toml_u64(key, item)?),
                "evictions_low_residence_duration_metric_threshold" => builder.evictions_low_residence_duration_metric_threshold(parse_toml_duration(key, item)?),
-                "disk_usage_based_eviction" => {
-                    tracing::info!("disk_usage_based_eviction: {:#?}", &item);
-                    builder.disk_usage_based_eviction(
-                    toml_edit::de::from_item(item.clone())
-                    .context("parse disk_usage_based_eviction")?)
-                },
                "ondemand_download_behavior_treat_error_as_warn" => builder.ondemand_download_behavior_treat_error_as_warn(parse_toml_bool(key, item)?),
                _ => bail!("unrecognized pageserver option '{key}'"),
            }
@@ -832,13 +808,6 @@ impl PageServerConf {
            );
        }

-        if let Some(item) = item.get("min_resident_size_override") {
-            t_conf.min_resident_size_override = Some(
-                toml_edit::de::from_item(item.clone())
-                    .context("parse min_resident_size_override")?,
-            );
-        }
-
        Ok(t_conf)
    }

@@ -881,7 +850,6 @@ impl PageServerConf {
                defaults::DEFAULT_EVICTIONS_LOW_RESIDENCE_DURATION_METRIC_THRESHOLD,
            )
            .unwrap(),
-            disk_usage_based_eviction: None,
            test_remote_failures: 0,
            ondemand_download_behavior_treat_error_as_warn: false,
        }
@@ -1090,7 +1058,6 @@ log_format = 'json'
                evictions_low_residence_duration_metric_threshold: humantime::parse_duration(
                    defaults::DEFAULT_EVICTIONS_LOW_RESIDENCE_DURATION_METRIC_THRESHOLD
                )?,
-                disk_usage_based_eviction: None,
                test_remote_failures: 0,
                ondemand_download_behavior_treat_error_as_warn: false,
            },
@@ -1145,7 +1112,6 @@ log_format = 'json'
                metric_collection_endpoint: Some(Url::parse("http://localhost:80/metrics")?),
                synthetic_size_calculation_interval: Duration::from_secs(333),
                evictions_low_residence_duration_metric_threshold: Duration::from_secs(444),
-                disk_usage_based_eviction: None,
                test_remote_failures: 0,
                ondemand_download_behavior_treat_error_as_warn: false,
            },
--- a/pageserver/src/disk_usage_eviction_task.rs
+++ b/pageserver/src/disk_usage_eviction_task.rs
@@ -1,728 +0,0 @@
-//! This module implements the pageserver-global disk-usage-based layer eviction task.
-//!
-//! # Mechanics
-//!
-//! Function `launch_disk_usage_global_eviction_task` starts a pageserver-global background
-//! loop that evicts layers in response to a shortage of available bytes
-//! in the $repo/tenants directory's filesystem.
-//!
-//! The loop runs periodically at a configurable `period`.
-//!
-//! Each loop iteration uses `statvfs` to determine filesystem-level space usage.
-//! It compares the returned usage data against two different types of thresholds.
-//! The iteration tries to evict layers until app-internal accounting says we should be below the thresholds.
-//! We cross-check this internal accounting with the real world by making another `statvfs` at the end of the iteration.
-//! We're good if that second statvfs shows that we're _actually_ below the configured thresholds.
-//! If we're still above one or more thresholds, we emit a warning log message, leaving it to the operator to investigate further.
-//!
-//! # Eviction Policy
-//!
-//! There are two thresholds:
-//! `max_usage_pct` is the relative available space, expressed in percent of the total filesystem space.
-//! If the actual usage is higher, the threshold is exceeded.
-//! `min_avail_bytes` is the absolute available space in bytes.
-//! If the actual usage is lower, the threshold is exceeded.
-//! If either of these thresholds is exceeded, the system is considered to have "disk pressure", and eviction
-//! is performed on the next iteration, to release disk space and bring the usage below the thresholds again.
-//! The iteration evicts layers in LRU fashion, but, with a weak reservation per tenant.
-//! The reservation is to keep the most recently accessed X bytes per tenant resident.
-//! If we cannot relieve pressure by evicting layers outside of the reservation, we
-//! start evicting layers that are part of the reservation, LRU first.
-//!
-//! The value for the per-tenant reservation is referred to as `tenant_min_resident_size`
-//! throughout the code, but, no actual variable carries that name.
-//! The per-tenant default value is the `max(tenant's layer file sizes, regardless of local or remote)`.
-//! The idea is to allow at least one layer to be resident per tenant, to ensure it can make forward progress
-//! during page reconstruction.
-//! An alternative default for all tenants can be specified in the `tenant_config` section of the config.
-//! Lastly, each tenant can have an override in their respective tenant config (`min_resident_size_override`).
-
-// Implementation notes:
-// - The `#[allow(dead_code)]` above various structs are to suppress warnings about only the Debug impl
-//   reading these fields. We use the Debug impl for semi-structured logging, though.
-
-use std::{
-    collections::HashMap,
-    path::Path,
-    sync::Arc,
-    time::{Duration, SystemTime},
-};
-
-use anyhow::Context;
-use remote_storage::GenericRemoteStorage;
-use serde::{Deserialize, Serialize};
-use tokio::time::Instant;
-use tokio_util::sync::CancellationToken;
-use tracing::{debug, error, info, instrument, warn, Instrument};
-use utils::serde_percent::Percent;
-
-use crate::{
-    config::PageServerConf,
-    task_mgr::{self, TaskKind, BACKGROUND_RUNTIME},
-    tenant::{self, storage_layer::PersistentLayer, Timeline},
-};
-
-#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
-pub struct DiskUsageEvictionTaskConfig {
-    pub max_usage_pct: Percent,
-    pub min_avail_bytes: u64,
-    #[serde(with = "humantime_serde")]
-    pub period: Duration,
-    #[cfg(feature = "testing")]
-    pub mock_statvfs: Option<crate::statvfs::mock::Behavior>,
-}
-
-#[derive(Default)]
-pub struct State {
-    /// Exclude http requests and background task from running at the same time.
-    mutex: tokio::sync::Mutex<()>,
-}
-
-pub fn launch_disk_usage_global_eviction_task(
-    conf: &'static PageServerConf,
-    storage: GenericRemoteStorage,
-    state: Arc<State>,
-) -> anyhow::Result<()> {
-    let Some(task_config) = &conf.disk_usage_based_eviction else {
-        info!("disk usage based eviction task not configured");
-        return Ok(());
-    };
-
-    info!("launching disk usage based eviction task");
-
-    task_mgr::spawn(
-        BACKGROUND_RUNTIME.handle(),
-        TaskKind::DiskUsageEviction,
-        None,
-        None,
-        "disk usage based eviction",
-        false,
-        async move {
-            disk_usage_eviction_task(
-                &state,
-                task_config,
-                storage,
-                &conf.tenants_path(),
-                task_mgr::shutdown_token(),
-            )
-            .await;
-            info!("disk usage based eviction task finishing");
-            Ok(())
-        },
-    );
-
-    Ok(())
-}
-
-#[instrument(skip_all)]
-async fn disk_usage_eviction_task(
-    state: &State,
-    task_config: &DiskUsageEvictionTaskConfig,
-    storage: GenericRemoteStorage,
-    tenants_dir: &Path,
-    cancel: CancellationToken,
-) {
-    use crate::tenant::tasks::random_init_delay;
-    {
-        if random_init_delay(task_config.period, &cancel)
-            .await
-            .is_err()
-        {
-            info!("shutting down");
-            return;
-        }
-    }
-
-    let mut iteration_no = 0;
-    loop {
-        iteration_no += 1;
-        let start = Instant::now();
-
-        async {
-            let res = disk_usage_eviction_task_iteration(
-                state,
-                task_config,
-                &storage,
-                tenants_dir,
-                &cancel,
-            )
-            .await;
-
-            match res {
-                Ok(()) => {}
-                Err(e) => {
-                    // these stat failures are expected to be very rare
-                    warn!("iteration failed, unexpected error: {e:#}");
-                }
-            }
-        }
-        .instrument(tracing::info_span!("iteration", iteration_no))
-        .await;
-
-        let sleep_until = start + task_config.period;
-        tokio::select! {
-            _ = tokio::time::sleep_until(sleep_until) => {},
-            _ = cancel.cancelled() => {
-                info!("shutting down");
-                break
-            }
-        }
-    }
-}
-
-pub trait Usage: Clone + Copy + std::fmt::Debug {
-    fn has_pressure(&self) -> bool;
-    fn add_available_bytes(&mut self, bytes: u64);
-}
-
-async fn disk_usage_eviction_task_iteration(
-    state: &State,
-    task_config: &DiskUsageEvictionTaskConfig,
-    storage: &GenericRemoteStorage,
-    tenants_dir: &Path,
-    cancel: &CancellationToken,
-) -> anyhow::Result<()> {
-    let usage_pre = filesystem_level_usage::get(tenants_dir, task_config)
-        .context("get filesystem-level disk usage before evictions")?;
-    let res = disk_usage_eviction_task_iteration_impl(state, storage, usage_pre, cancel).await;
-    match res {
-        Ok(outcome) => {
-            debug!(?outcome, "disk_usage_eviction_iteration finished");
-            match outcome {
-                IterationOutcome::NoPressure | IterationOutcome::Cancelled => {
-                    // nothing to do, select statement below will handle things
-                }
-                IterationOutcome::Finished(outcome) => {
-                    // Verify with statvfs whether we made any real progress
-                    let after = filesystem_level_usage::get(tenants_dir, task_config)
-                        // It's quite unlikely to hit the error here. Keep the code simple and bail out.
-                        .context("get filesystem-level disk usage after evictions")?;
-
-                    debug!(?after, "disk usage");
-
-                    if after.has_pressure() {
-                        // Don't bother doing an out-of-order iteration here now.
-                        // In practice, the task period is set to a value in the tens-of-seconds range,
-                        // which will cause another iteration to happen soon enough.
-                        // TODO: deltas between the three different usages would be helpful,
-                        // consider MiB, GiB, TiB
-                        warn!(?outcome, ?after, "disk usage still high");
-                    } else {
-                        info!(?outcome, ?after, "disk usage pressure relieved");
-                    }
-                }
-            }
-        }
-        Err(e) => {
-            error!("disk_usage_eviction_iteration failed: {:#}", e);
-        }
-    }
-
-    Ok(())
-}
-
-#[derive(Debug, Serialize)]
-#[allow(clippy::large_enum_variant)]
-pub enum IterationOutcome<U> {
-    NoPressure,
-    Cancelled,
-    Finished(IterationOutcomeFinished<U>),
-}
-
-#[allow(dead_code)]
-#[derive(Debug, Serialize)]
-pub struct IterationOutcomeFinished<U> {
-    /// The actual usage observed before we started the iteration.
-    before: U,
-    /// The expected value for `after`, according to internal accounting, after phase 1.
-    planned: PlannedUsage<U>,
-    /// The outcome of phase 2, where we actually do the evictions.
-    ///
-    /// If all layers that phase 1 planned to evict _can_ actually get evicted, this will
-    /// be the same as `planned`.
-    assumed: AssumedUsage<U>,
-}
-
-#[derive(Debug, Serialize)]
-#[allow(dead_code)]
-struct AssumedUsage<U> {
-    /// The expected value for `after`, after phase 2.
-    projected_after: U,
-    /// The layers we failed to evict during phase 2.
-    failed: LayerCount,
-}
-
-#[allow(dead_code)]
-#[derive(Debug, Serialize)]
-struct PlannedUsage<U> {
-    respecting_tenant_min_resident_size: U,
-    fallback_to_global_lru: Option<U>,
-}
-
-#[allow(dead_code)]
-#[derive(Debug, Default, Serialize)]
-struct LayerCount {
-    file_sizes: u64,
-    count: usize,
-}
-
-pub async fn disk_usage_eviction_task_iteration_impl<U: Usage>(
-    state: &State,
-    storage: &GenericRemoteStorage,
-    usage_pre: U,
-    cancel: &CancellationToken,
-) -> anyhow::Result<IterationOutcome<U>> {
-    // use tokio's mutex to get a Sync guard (instead of std::sync::Mutex)
-    let _g = state
-        .mutex
-        .try_lock()
-        .map_err(|_| anyhow::anyhow!("iteration is already executing"))?;
-
-    debug!(?usage_pre, "disk usage");
-
-    if !usage_pre.has_pressure() {
-        return Ok(IterationOutcome::NoPressure);
-    }
-
-    warn!(
-        ?usage_pre,
-        "running disk usage based eviction due to pressure"
-    );
-
-    let candidates = match collect_eviction_candidates(cancel).await? {
-        EvictionCandidates::Cancelled => {
-            return Ok(IterationOutcome::Cancelled);
-        }
-        EvictionCandidates::Finished(partitioned) => partitioned,
-    };
-
-    // Debug-log the list of candidates
-    let now = SystemTime::now();
-    for (i, (partition, candidate)) in candidates.iter().enumerate() {
-        debug!(
-            "cand {}/{}: size={}, no_access_for={}us, parition={:?}, tenant={} timeline={} layer={}",
-            i + 1,
-            candidates.len(),
-            candidate.layer.file_size(),
-            now.duration_since(candidate.last_activity_ts)
-                .unwrap()
-                .as_micros(),
-            partition,
-            candidate.layer.get_tenant_id(),
-            candidate.layer.get_timeline_id(),
-            candidate.layer.filename().file_name(),
-        );
-    }
-
-    // phase1: select victims to relieve pressure
-    //
-    // Walk through the list of candidates, until we have accumulated enough layers to get
-    // us back under the pressure threshold. 'usage_planned' is updated so that it tracks
-    // how much disk space would be used after evicting all the layers up to the current
-    // point in the list. The layers are collected in 'batched', grouped per timeline.
-    //
-    // If we get far enough in the list that we start to evict layers that are below
-    // the tenant's min-resident-size threshold, print a warning, and memorize the disk
-    // usage at that point, in 'usage_planned_min_resident_size_respecting'.
-    let mut batched: HashMap<_, Vec<Arc<dyn PersistentLayer>>> = HashMap::new();
-    let mut warned = None;
-    let mut usage_planned = usage_pre;
-    for (i, (partition, candidate)) in candidates.into_iter().enumerate() {
-        if !usage_planned.has_pressure() {
-            debug!(
-                no_candidates_evicted = i,
-                "took enough candidates for pressure to be relieved"
-            );
-            break;
-        }
-
-        if partition == MinResidentSizePartition::Below && warned.is_none() {
-            warn!(?usage_pre, ?usage_planned, candidate_no=i, "tenant_min_resident_size-respecting LRU would not relieve pressure, evicting more following global LRU policy");
-            warned = Some(usage_planned);
-        }
-
-        usage_planned.add_available_bytes(candidate.layer.file_size());
-
-        batched
-            .entry(TimelineKey(candidate.timeline))
-            .or_default()
-            .push(candidate.layer);
-    }
-
-    let usage_planned = match warned {
-        Some(respecting_tenant_min_resident_size) => PlannedUsage {
-            respecting_tenant_min_resident_size,
-            fallback_to_global_lru: Some(usage_planned),
-        },
-        None => PlannedUsage {
-            respecting_tenant_min_resident_size: usage_planned,
-            fallback_to_global_lru: None,
-        },
-    };
-    debug!(?usage_planned, "usage planned");
-
-    // phase2: evict victims batched by timeline
-
-    // After the loop, `usage_assumed` is the post-eviction usage,
-    // according to internal accounting.
-    let mut usage_assumed = usage_pre;
-    let mut evictions_failed = LayerCount::default();
-    for (timeline, batch) in batched {
-        let tenant_id = timeline.tenant_id;
-        let timeline_id = timeline.timeline_id;
-        let batch_size = batch.len();
-
-        debug!(%timeline_id, "evicting batch for timeline");
-
-        async {
-            let results = timeline.evict_layers(storage, &batch, cancel.clone()).await;
-
-            match results {
-                Err(e) => {
-                    warn!("failed to evict batch: {:#}", e);
-                }
-                Ok(results) => {
-                    assert_eq!(results.len(), batch.len());
-                    for (result, layer) in results.into_iter().zip(batch.iter()) {
-                        match result {
-                            Some(Ok(true)) => {
-                                usage_assumed.add_available_bytes(layer.file_size());
-                            }
-                            Some(Ok(false)) => {
-                                // this is:
-                                // - Replacement::{NotFound, Unexpected}
-                                // - it cannot be is_remote_layer, filtered already
-                                evictions_failed.file_sizes += layer.file_size();
-                                evictions_failed.count += 1;
-                            }
-                            None => {
-                                assert!(cancel.is_cancelled());
-                                return;
-                            }
-                            Some(Err(e)) => {
-                                // we really shouldn't be getting this, precondition failure
-                                error!("failed to evict layer: {:#}", e);
-                            }
-                        }
-                    }
-                }
-            }
-        }
-        .instrument(tracing::info_span!("evict_batch", %tenant_id, %timeline_id, batch_size))
-        .await;
-
-        if cancel.is_cancelled() {
-            return Ok(IterationOutcome::Cancelled);
-        }
-    }
-
-    Ok(IterationOutcome::Finished(IterationOutcomeFinished {
-        before: usage_pre,
-        planned: usage_planned,
-        assumed: AssumedUsage {
-            projected_after: usage_assumed,
-            failed: evictions_failed,
-        },
-    }))
-}
-
-#[derive(Clone)]
-struct EvictionCandidate {
-    timeline: Arc<Timeline>,
-    layer: Arc<dyn PersistentLayer>,
-    last_activity_ts: SystemTime,
-}
-
-#[derive(Debug, Clone, Copy, PartialEq, Eq, PartialOrd, Ord)]
-enum MinResidentSizePartition {
-    Above,
-    Below,
-}
-
-enum EvictionCandidates {
-    Cancelled,
-    Finished(Vec<(MinResidentSizePartition, EvictionCandidate)>),
-}
-
-/// Gather the eviction candidates.
-///
-/// The returned `Ok(EvictionCandidates::Finished(candidates))` is sorted in eviction
-/// order. A caller that evicts in that order, until pressure is relieved, implements
-/// the eviction policy outlined in the module comment.
-///
-/// # Example
-///
-/// Imagine that there are two tenants, A and B, with five layers each, a-e.
-/// Each layer has size 100, and both tenant's min_resident_size is 150.
-/// The eviction order would be
-///
-/// ```text
-/// partition last_activity_ts    tenant/layer
-/// Above     18:30               A/c
-/// Above     19:00               A/b
-/// Above     18:29               B/c
-/// Above     19:05               B/b
-/// Above     20:00               B/a
-/// Above     20:03               A/a
-/// Below     20:30               A/d
-/// Below     20:40               B/d
-/// Below     20:45               B/e
-/// Below     20:58               A/e
-/// ```
-///
-/// Now, if we need to evict 300 bytes to relieve pressure, we'd evict `A/c, A/b, B/c`.
-/// They are all in the `Above` partition, so, we respected each tenant's min_resident_size.
-///
-/// But, if we need to evict 900 bytes to relieve pressure, we'd evict
-/// `A/c, A/b, B/c, B/b, B/a, A/a, A/d, B/d, B/e`, reaching into the `Below` partition
-/// after exhauting the `Above` partition.
-/// So, we did not respect each tenant's min_resident_size.
-async fn collect_eviction_candidates(
-    cancel: &CancellationToken,
-) -> anyhow::Result<EvictionCandidates> {
-    // get a snapshot of the list of tenants
-    let tenants = tenant::mgr::list_tenants()
-        .await
-        .context("get list of tenants")?;
-
-    let mut candidates = Vec::new();
-
-    for (tenant_id, _state) in &tenants {
-        if cancel.is_cancelled() {
-            return Ok(EvictionCandidates::Cancelled);
-        }
-        let tenant = match tenant::mgr::get_tenant(*tenant_id, true).await {
-            Ok(tenant) => tenant,
-            Err(e) => {
-                // this can happen if tenant has lifecycle transition after we fetched it
-                debug!("failed to get tenant: {e:#}");
-                continue;
-            }
-        };
-
-        // collect layers from all timelines in this tenant
-        //
-        // If one of the timelines becomes `!is_active()` during the iteration,
-        // for example because we're shutting down, then `max_layer_size` can be too small.
-        // That's OK. This code only runs under a disk pressure situation, and being
-        // a little unfair to tenants during shutdown in such a situation is tolerable.
-        let mut tenant_candidates = Vec::new();
-        let mut max_layer_size = 0;
-        for tl in tenant.list_timelines() {
-            if !tl.is_active() {
-                continue;
-            }
-            let info = tl.get_local_layers_for_disk_usage_eviction();
-            debug!(tenant_id=%tl.tenant_id, timeline_id=%tl.timeline_id, "timeline resident layers count: {}", info.resident_layers.len());
-            tenant_candidates.extend(
-                info.resident_layers
-                    .into_iter()
-                    .map(|layer_infos| (tl.clone(), layer_infos)),
-            );
-            max_layer_size = max_layer_size.max(info.max_layer_size.unwrap_or(0));
-
-            if cancel.is_cancelled() {
-                return Ok(EvictionCandidates::Cancelled);
-            }
-        }
-
-        // `min_resident_size` defaults to maximum layer file size of the tenant.
-        // This ensures that each tenant can have at least one layer resident at a given time,
-        // ensuring forward progress for a single Timeline::get in that tenant.
-        // It's a questionable heuristic since, usually, there are many Timeline::get
-        // requests going on for a tenant, and, at least in Neon prod, the median
-        // layer file size is much smaller than the compaction target size.
-        // We could be better here, e.g., sum of all L0 layers + most recent L1 layer.
-        // That's what's typically used by the various background loops.
-        //
-        // The default can be overriden with a fixed value in the tenant conf.
-        // A default override can be put in the default tenant conf in the pageserver.toml.
-        let min_resident_size = if let Some(s) = tenant.get_min_resident_size_override() {
-            debug!(
-                tenant_id=%tenant.tenant_id(),
-                overriden_size=s,
-                "using overridden min resident size for tenant"
-            );
-            s
-        } else {
-            debug!(
-                tenant_id=%tenant.tenant_id(),
-                max_layer_size,
-                "using max layer size as min_resident_size for tenant",
-            );
-            max_layer_size
-        };
-
-        // Sort layers most-recently-used first, then partition by
-        // cumsum above/below min_resident_size.
-        tenant_candidates
-            .sort_unstable_by_key(|(_, layer_info)| std::cmp::Reverse(layer_info.last_activity_ts));
-        let mut cumsum: i128 = 0;
-        for (timeline, layer_info) in tenant_candidates.into_iter() {
-            let file_size = layer_info.file_size();
-            let candidate = EvictionCandidate {
-                timeline,
-                last_activity_ts: layer_info.last_activity_ts,
-                layer: layer_info.layer,
-            };
-            let partition = if cumsum > min_resident_size as i128 {
-                MinResidentSizePartition::Above
-            } else {
-                MinResidentSizePartition::Below
-            };
-            candidates.push((partition, candidate));
-            cumsum += i128::from(file_size);
-        }
-    }
-
-    debug_assert!(MinResidentSizePartition::Above < MinResidentSizePartition::Below,
-        "as explained in the function's doc comment, layers that aren't in the tenant's min_resident_size are evicted first");
-    candidates
-        .sort_unstable_by_key(|(partition, candidate)| (*partition, candidate.last_activity_ts));
-
-    Ok(EvictionCandidates::Finished(candidates))
-}
-
-struct TimelineKey(Arc<Timeline>);
-
-impl PartialEq for TimelineKey {
-    fn eq(&self, other: &Self) -> bool {
-        Arc::ptr_eq(&self.0, &other.0)
-    }
-}
-
-impl Eq for TimelineKey {}
-
-impl std::hash::Hash for TimelineKey {
-    fn hash<H: std::hash::Hasher>(&self, state: &mut H) {
-        Arc::as_ptr(&self.0).hash(state);
-    }
-}
-
-impl std::ops::Deref for TimelineKey {
-    type Target = Timeline;
-
-    fn deref(&self) -> &Self::Target {
-        self.0.as_ref()
-    }
-}
-
-mod filesystem_level_usage {
-    use std::path::Path;
-
-    use anyhow::Context;
-
-    use crate::statvfs::Statvfs;
-
-    use super::DiskUsageEvictionTaskConfig;
-
-    #[derive(Debug, Clone, Copy)]
-    #[allow(dead_code)]
-    pub struct Usage<'a> {
-        config: &'a DiskUsageEvictionTaskConfig,
-
-        /// Filesystem capacity
-        total_bytes: u64,
-        /// Free filesystem space
-        avail_bytes: u64,
-    }
-
-    impl super::Usage for Usage<'_> {
-        fn has_pressure(&self) -> bool {
-            let usage_pct =
-                (100.0 * (1.0 - ((self.avail_bytes as f64) / (self.total_bytes as f64)))) as u64;
-
-            let pressures = [
-                (
-                    "min_avail_bytes",
-                    self.avail_bytes < self.config.min_avail_bytes,
-                ),
-                (
-                    "max_usage_pct",
-                    usage_pct >= self.config.max_usage_pct.get() as u64,
-                ),
-            ];
-
-            pressures.into_iter().any(|(_, has_pressure)| has_pressure)
-        }
-
-        fn add_available_bytes(&mut self, bytes: u64) {
-            self.avail_bytes += bytes;
-        }
-    }
-
-    pub fn get<'a>(
-        tenants_dir: &Path,
-        config: &'a DiskUsageEvictionTaskConfig,
-    ) -> anyhow::Result<Usage<'a>> {
-        let mock_config = {
-            #[cfg(feature = "testing")]
-            {
-                config.mock_statvfs.as_ref()
-            }
-            #[cfg(not(feature = "testing"))]
-            {
-                None
-            }
-        };
-
-        let stat = Statvfs::get(tenants_dir, mock_config)
-            .context("statvfs failed, presumably directory got unlinked")?;
-
-        // https://unix.stackexchange.com/a/703650
-        let blocksize = if stat.fragment_size() > 0 {
-            stat.fragment_size()
-        } else {
-            stat.block_size()
-        };
-
-        // use blocks_available (b_avail) since, pageserver runs as unprivileged user
-        let avail_bytes = stat.blocks_available() * blocksize;
-        let total_bytes = stat.blocks() * blocksize;
-
-        Ok(Usage {
-            config,
-            total_bytes,
-            avail_bytes,
-        })
-    }
-
-    #[test]
-    fn max_usage_pct_pressure() {
-        use super::Usage as _;
-        use std::time::Duration;
-        use utils::serde_percent::Percent;
-
-        let mut usage = Usage {
-            config: &DiskUsageEvictionTaskConfig {
-                max_usage_pct: Percent::new(85).unwrap(),
-                min_avail_bytes: 0,
-                period: Duration::MAX,
-                #[cfg(feature = "testing")]
-                mock_statvfs: None,
-            },
-            total_bytes: 100_000,
-            avail_bytes: 0,
-        };
-
-        assert!(usage.has_pressure(), "expected pressure at 100%");
-
-        usage.add_available_bytes(14_000);
-        assert!(usage.has_pressure(), "expected pressure at 86%");
-
-        usage.add_available_bytes(999);
-        assert!(usage.has_pressure(), "expected pressure at 85.001%");
-
-        usage.add_available_bytes(1);
-        assert!(usage.has_pressure(), "expected pressure at precisely 85%");
-
-        usage.add_available_bytes(1);
-        assert!(!usage.has_pressure(), "no pressure at 84.999%");
-
-        usage.add_available_bytes(999);
-        assert!(!usage.has_pressure(), "no pressure at 84%");
-
-        usage.add_available_bytes(16_000);
-        assert!(!usage.has_pressure());
-    }
-}
--- a/pageserver/src/http/openapi_spec.yml
+++ b/pageserver/src/http/openapi_spec.yml
@@ -27,31 +27,6 @@ paths:
                  id:
                    type: integer

-  /v1/disk_usage_eviction/run:
-    put:
-      description: Do an iteration of disk-usage-based eviction to evict a given amount of disk space.
-      security: []
-      requestBody:
-        content:
-          application/json:
-            schema:
-              type: object
-              required:
-                - evict_bytes
-              properties:
-                evict_bytes:
-                  type: integer
-      responses:
-        "200":
-          description: |
-            The run completed.
-            This does not necessarily mean that we actually evicted `evict_bytes`.
-            Examine the returned object for detail, or, just watch the actual effect of the call using `du` or `df`.
-          content:
-            application/json:
-              schema:
-                type: object
-
  /v1/tenant/{tenant_id}:
    parameters:
      - name: tenant_id
@@ -829,9 +804,12 @@ components:
      type: object
      required:
        - id
+        - state
      properties:
        id:
          type: string
+        state:
+          type: string
        current_physical_size:
          type: integer
        has_in_progress_downloads:
--- a/pageserver/src/http/routes.rs
+++ b/pageserver/src/http/routes.rs
@@ -18,7 +18,6 @@ use super::models::{
    TimelineCreateRequest, TimelineGcRequest, TimelineInfo,
 };
 use crate::context::{DownloadBehavior, RequestContext};
-use crate::disk_usage_eviction_task;
 use crate::pgdatadir_mapping::LsnForTimestamp;
 use crate::task_mgr::TaskKind;
 use crate::tenant::config::TenantConfOpt;
@@ -49,7 +48,6 @@ struct State {
    auth: Option<Arc<JwtAuth>>,
    allowlist_routes: Vec<Uri>,
    remote_storage: Option<GenericRemoteStorage>,
-    disk_usage_eviction_state: Arc<disk_usage_eviction_task::State>,
 }

 impl State {
@@ -57,7 +55,6 @@ impl State {
        conf: &'static PageServerConf,
        auth: Option<Arc<JwtAuth>>,
        remote_storage: Option<GenericRemoteStorage>,
-        disk_usage_eviction_state: Arc<disk_usage_eviction_task::State>,
    ) -> anyhow::Result<Self> {
        let allowlist_routes = ["/v1/status", "/v1/doc", "/swagger.yml"]
            .iter()
@@ -68,7 +65,6 @@ impl State {
            auth,
            allowlist_routes,
            remote_storage,
-            disk_usage_eviction_state,
        })
    }
 }
@@ -465,7 +461,7 @@ async fn tenant_list_handler(request: Request<Body>) -> Result<Response<Body>, A
        .iter()
        .map(|(id, state)| TenantInfo {
            id: *id,
-            state: state.clone(),
+            state: *state,
            current_physical_size: None,
            has_in_progress_downloads: Some(state.has_in_progress_downloads()),
        })
@@ -490,7 +486,7 @@ async fn tenant_status(request: Request<Body>) -> Result<Response<Body>, ApiErro
        let state = tenant.current_state();
        Ok(TenantInfo {
            id: tenant_id,
-            state: state.clone(),
+            state,
            current_physical_size: Some(current_physical_size),
            has_in_progress_downloads: Some(state.has_in_progress_downloads()),
        })
@@ -779,8 +775,6 @@ async fn tenant_create_handler(mut request: Request<Body>) -> Result<Response<Bo
        );
    }

-    tenant_conf.min_resident_size_override = request_data.min_resident_size_override;
-
    let target_tenant_id = request_data
        .new_tenant_id
        .map(TenantId::from)
@@ -912,8 +906,6 @@ async fn update_tenant_config_handler(
        );
    }

-    tenant_conf.min_resident_size_override = request_data.min_resident_size_override;
-
    let state = get_state(&request);
    mgr::set_new_tenant_config(state.conf, tenant_conf, tenant_id)
        .instrument(info_span!("tenant_config", tenant = ?tenant_id))
@@ -922,20 +914,6 @@ async fn update_tenant_config_handler(
    json_response(StatusCode::OK, ())
 }

-/// Testing helper to transition a tenant to [`crate::tenant::TenantState::Broken`].
-#[cfg(feature = "testing")]
-async fn handle_tenant_break(r: Request<Body>) -> Result<Response<Body>, ApiError> {
-    let tenant_id: TenantId = parse_request_param(&r, "tenant_id")?;
-
-    let tenant = crate::tenant::mgr::get_tenant(tenant_id, true)
-        .await
-        .map_err(|_| ApiError::Conflict(String::from("no active tenant found")))?;
-
-    tenant.set_broken("broken from test".to_owned());
-
-    json_response(StatusCode::OK, ())
-}
-
 #[cfg(feature = "testing")]
 async fn failpoints_handler(mut request: Request<Body>) -> Result<Response<Body>, ApiError> {
    if !fail::has_failpoints() {
@@ -1085,89 +1063,6 @@ async fn always_panic_handler(req: Request<Body>) -> Result<Response<Body>, ApiE
    json_response(StatusCode::NO_CONTENT, ())
 }

-async fn disk_usage_eviction_run(mut r: Request<Body>) -> Result<Response<Body>, ApiError> {
-    check_permission(&r, None)?;
-
-    #[derive(Debug, Clone, Copy, serde::Serialize, serde::Deserialize)]
-    struct Config {
-        /// How many bytes to evict before reporting that pressure is relieved.
-        evict_bytes: u64,
-    }
-
-    #[derive(Debug, Clone, Copy, serde::Serialize)]
-    struct Usage {
-        // remains unchanged after instantiation of the struct
-        config: Config,
-        // updated by `add_available_bytes`
-        freed_bytes: u64,
-    }
-
-    impl crate::disk_usage_eviction_task::Usage for Usage {
-        fn has_pressure(&self) -> bool {
-            self.config.evict_bytes > self.freed_bytes
-        }
-
-        fn add_available_bytes(&mut self, bytes: u64) {
-            self.freed_bytes += bytes;
-        }
-    }
-
-    let config = json_request::<Config>(&mut r)
-        .await
-        .map_err(|_| ApiError::BadRequest(anyhow::anyhow!("invalid JSON body")))?;
-
-    let usage = Usage {
-        config,
-        freed_bytes: 0,
-    };
-
-    use crate::task_mgr::MGMT_REQUEST_RUNTIME;
-
-    let (tx, rx) = tokio::sync::oneshot::channel();
-
-    let state = get_state(&r);
-
-    let Some(storage) = state.remote_storage.clone() else {
-        return Err(ApiError::InternalServerError(anyhow::anyhow!(
-            "remote storage not configured, cannot run eviction iteration"
-        )))
-    };
-
-    let state = state.disk_usage_eviction_state.clone();
-
-    let cancel = CancellationToken::new();
-    let child_cancel = cancel.clone();
-    let _g = cancel.drop_guard();
-
-    crate::task_mgr::spawn(
-        MGMT_REQUEST_RUNTIME.handle(),
-        TaskKind::DiskUsageEviction,
-        None,
-        None,
-        "ondemand disk usage eviction",
-        false,
-        async move {
-            let res = crate::disk_usage_eviction_task::disk_usage_eviction_task_iteration_impl(
-                &state,
-                &storage,
-                usage,
-                &child_cancel,
-            )
-            .await;
-
-            info!(?res, "disk_usage_eviction_task_iteration_impl finished");
-
-            let _ = tx.send(res);
-            Ok(())
-        }
-        .in_current_span(),
-    );
-
-    let response = rx.await.unwrap().map_err(ApiError::InternalServerError)?;
-
-    json_response(StatusCode::OK, response)
-}
-
 async fn handler_404(_: Request<Body>) -> Result<Response<Body>, ApiError> {
    json_response(
        StatusCode::NOT_FOUND,
@@ -1180,7 +1075,6 @@ pub fn make_router(
    launch_ts: &'static LaunchTimestamp,
    auth: Option<Arc<JwtAuth>>,
    remote_storage: Option<GenericRemoteStorage>,
-    disk_usage_eviction_state: Arc<disk_usage_eviction_task::State>,
 ) -> anyhow::Result<RouterBuilder<hyper::Body, ApiError>> {
    let spec = include_bytes!("openapi_spec.yml");
    let mut router = attach_openapi_ui(endpoint::make_router(), spec, "/swagger.yml", "/v1/doc");
@@ -1225,8 +1119,7 @@ pub fn make_router(

    Ok(router
        .data(Arc::new(
-            State::new(conf, auth, remote_storage, disk_usage_eviction_state)
-                .context("Failed to initialize router state")?,
+            State::new(conf, auth, remote_storage).context("Failed to initialize router state")?,
        ))
        .get("/v1/status", |r| RequestSpan(status_handler).handle(r))
        .put(
@@ -1307,13 +1200,6 @@ pub fn make_router(
            "/v1/tenant/:tenant_id/timeline/:timeline_id/layer/:layer_file_name",
            |r| RequestSpan(evict_timeline_layer_handler).handle(r),
        )
-        .put("/v1/disk_usage_eviction/run", |r| {
-            RequestSpan(disk_usage_eviction_run).handle(r)
-        })
-        .put(
-            "/v1/tenant/:tenant_id/break",
-            testing_api!("set tenant state to broken", handle_tenant_break),
-        )
        .get("/v1/panic", |r| RequestSpan(always_panic_handler).handle(r))
        .any(handler_404))
 }
--- a/pageserver/src/lib.rs
+++ b/pageserver/src/lib.rs
@@ -4,7 +4,6 @@ pub mod broker_client;
 pub mod config;
 pub mod consumption_metrics;
 pub mod context;
-pub mod disk_usage_eviction_task;
 pub mod http;
 pub mod import_datadir;
 pub mod keyspace;
@@ -13,7 +12,6 @@ pub mod page_cache;
 pub mod page_service;
 pub mod pgdatadir_mapping;
 pub mod repository;
-pub(crate) mod statvfs;
 pub mod task_mgr;
 pub mod tenant;
 pub mod trace;
--- a/pageserver/src/metrics.rs
+++ b/pageserver/src/metrics.rs
@@ -6,8 +6,7 @@ use metrics::{
    UIntGauge, UIntGaugeVec,
 };
 use once_cell::sync::Lazy;
-use pageserver_api::models::TenantState;
-use strum::VariantNames;
+use pageserver_api::models::state;
 use utils::id::{TenantId, TimelineId};

 /// Prometheus histogram buckets (in seconds) for operations in the critical
@@ -148,6 +147,15 @@ static CURRENT_LOGICAL_SIZE: Lazy<UIntGaugeVec> = Lazy::new(|| {
    .expect("failed to define current logical size metric")
 });

+// Metrics collected on tenant states.
+const TENANT_STATE_OPTIONS: &[&str] = &[
+    state::LOADING,
+    state::ATTACHING,
+    state::ACTIVE,
+    state::STOPPING,
+    state::BROKEN,
+];
+
 pub static TENANT_STATE_METRIC: Lazy<UIntGaugeVec> = Lazy::new(|| {
    register_uint_gauge_vec!(
        "pageserver_tenant_states_count",
@@ -249,7 +257,7 @@ impl EvictionsWithLowResidenceDuration {
    }

    pub fn observe(&self, observed_value: Duration) {
-        if observed_value < self.threshold {
+        if self.threshold < observed_value {
            self.counter
                .as_ref()
                .expect("nobody calls this function after `remove_from_vec`")
@@ -699,7 +707,7 @@ impl Drop for TimelineMetrics {
 pub fn remove_tenant_metrics(tenant_id: &TenantId) {
    let tid = tenant_id.to_string();
    let _ = TENANT_SYNTHETIC_SIZE_METRIC.remove_label_values(&[&tid]);
-    for state in TenantState::VARIANTS {
+    for state in TENANT_STATE_OPTIONS {
        let _ = TENANT_STATE_METRIC.remove_label_values(&[&tid, state]);
    }
 }
--- a/pageserver/src/statvfs.rs
+++ b/pageserver/src/statvfs.rs
@@ -1,150 +0,0 @@
-//! Wrapper around nix::sys::statvfs::Statvfs that allows for mocking.
-
-use std::path::Path;
-
-pub enum Statvfs {
-    Real(nix::sys::statvfs::Statvfs),
-    Mock(mock::Statvfs),
-}
-
-// NB: on macOS, the block count type of struct statvfs is u32.
-// The workaround seems to be to use the non-standard statfs64 call.
-// Sincce it should only be a problem on > 2TiB disks, let's ignore
-// the problem for now and upcast to u64.
-impl Statvfs {
-    pub fn get(tenants_dir: &Path, mocked: Option<&mock::Behavior>) -> nix::Result<Self> {
-        if let Some(mocked) = mocked {
-            Ok(Statvfs::Mock(mock::get(tenants_dir, mocked)?))
-        } else {
-            Ok(Statvfs::Real(nix::sys::statvfs::statvfs(tenants_dir)?))
-        }
-    }
-
-    // NB: allow() because the block count type is u32 on macOS.
-    #[allow(clippy::useless_conversion)]
-    pub fn blocks(&self) -> u64 {
-        match self {
-            Statvfs::Real(stat) => u64::try_from(stat.blocks()).unwrap(),
-            Statvfs::Mock(stat) => stat.blocks,
-        }
-    }
-
-    // NB: allow() because the block count type is u32 on macOS.
-    #[allow(clippy::useless_conversion)]
-    pub fn blocks_available(&self) -> u64 {
-        match self {
-            Statvfs::Real(stat) => u64::try_from(stat.blocks_available()).unwrap(),
-            Statvfs::Mock(stat) => stat.blocks_available,
-        }
-    }
-
-    pub fn fragment_size(&self) -> u64 {
-        match self {
-            Statvfs::Real(stat) => stat.fragment_size(),
-            Statvfs::Mock(stat) => stat.fragment_size,
-        }
-    }
-
-    pub fn block_size(&self) -> u64 {
-        match self {
-            Statvfs::Real(stat) => stat.block_size(),
-            Statvfs::Mock(stat) => stat.block_size,
-        }
-    }
-}
-
-pub mod mock {
-    use anyhow::Context;
-    use regex::Regex;
-    use std::path::Path;
-    use tracing::log::info;
-
-    #[derive(Debug, Clone, PartialEq, Eq, serde::Serialize, serde::Deserialize)]
-    #[serde(tag = "type")]
-    pub enum Behavior {
-        Success {
-            blocksize: u64,
-            total_blocks: u64,
-            name_filter: Option<utils::serde_regex::Regex>,
-        },
-        Failure {
-            mocked_error: MockedError,
-        },
-    }
-
-    #[derive(Debug, Clone, Copy, PartialEq, Eq, serde::Serialize, serde::Deserialize)]
-    #[allow(clippy::upper_case_acronyms)]
-    pub enum MockedError {
-        EIO,
-    }
-
-    impl From<MockedError> for nix::Error {
-        fn from(e: MockedError) -> Self {
-            match e {
-                MockedError::EIO => nix::Error::EIO,
-            }
-        }
-    }
-
-    pub fn get(tenants_dir: &Path, behavior: &Behavior) -> nix::Result<Statvfs> {
-        info!("running mocked statvfs");
-
-        match behavior {
-            Behavior::Success {
-                blocksize,
-                total_blocks,
-                ref name_filter,
-            } => {
-                let used_bytes = walk_dir_disk_usage(tenants_dir, name_filter.as_deref()).unwrap();
-
-                // round it up to the nearest block multiple
-                let used_blocks = (used_bytes + (blocksize - 1)) / blocksize;
-
-                if used_blocks > *total_blocks {
-                    panic!(
-                        "mocking error: used_blocks > total_blocks: {used_blocks} > {total_blocks}"
-                    );
-                }
-
-                let avail_blocks = total_blocks - used_blocks;
-
-                Ok(Statvfs {
-                    blocks: *total_blocks,
-                    blocks_available: avail_blocks,
-                    fragment_size: *blocksize,
-                    block_size: *blocksize,
-                })
-            }
-            Behavior::Failure { mocked_error } => Err((*mocked_error).into()),
-        }
-    }
-
-    fn walk_dir_disk_usage(path: &Path, name_filter: Option<&Regex>) -> anyhow::Result<u64> {
-        let mut total = 0;
-        for entry in walkdir::WalkDir::new(path) {
-            let entry = entry?;
-            if !entry.file_type().is_file() {
-                continue;
-            }
-            if !name_filter
-                .as_ref()
-                .map(|filter| filter.is_match(entry.file_name().to_str().unwrap()))
-                .unwrap_or(true)
-            {
-                continue;
-            }
-            total += entry
-                .metadata()
-                .with_context(|| format!("get metadata of {:?}", entry.path()))?
-                .len();
-        }
-        Ok(total)
-    }
-
-    pub struct Statvfs {
-        pub blocks: u64,
-        pub blocks_available: u64,
-        pub fragment_size: u64,
-        pub block_size: u64,
-    }
-}
--- a/pageserver/src/task_mgr.rs
+++ b/pageserver/src/task_mgr.rs
@@ -234,9 +234,6 @@ pub enum TaskKind {
    // Eviction. One per timeline.
    Eviction,

-    /// See [`crate::disk_usage_eviction_task`].
-    DiskUsageEviction,
-
    // Initial logical size calculation
    InitialLogicalSizeCalculation,

--- a/pageserver/src/tenant.rs
+++ b/pageserver/src/tenant.rs
@@ -95,7 +95,7 @@ mod timeline;

 pub mod size;

-pub use timeline::{LocalLayerInfoForDiskUsageEviction, PageReconstructError, Timeline};
+pub use timeline::{PageReconstructError, Timeline};

 // re-export this function so that page_cache.rs can use it.
 pub use crate::tenant::ephemeral_file::writeback as writeback_ephemeral_file;
@@ -177,9 +177,9 @@ impl UninitializedTimeline<'_> {
    ///
    /// The new timeline is initialized in Active state, and its background jobs are
    /// started
-    pub fn initialize(self, ctx: &RequestContext) -> anyhow::Result<Arc<Timeline>> {
+    pub fn initialize(self, _ctx: &RequestContext) -> anyhow::Result<Arc<Timeline>> {
        let mut timelines = self.owning_tenant.timelines.lock().unwrap();
-        self.initialize_with_lock(ctx, &mut timelines, true, true)
+        self.initialize_with_lock(&mut timelines, true, true)
    }

    /// Like `initialize`, but the caller is already holding lock on Tenant::timelines.
@@ -189,7 +189,6 @@ impl UninitializedTimeline<'_> {
    /// been initialized.
    fn initialize_with_lock(
        mut self,
-        ctx: &RequestContext,
        timelines: &mut HashMap<TimelineId, Arc<Timeline>>,
        load_layer_map: bool,
        activate: bool,
@@ -230,9 +229,7 @@ impl UninitializedTimeline<'_> {
                new_timeline.maybe_spawn_flush_loop();

                if activate {
-                    new_timeline
-                        .activate(ctx)
-                        .context("initializing timeline activation")?;
+                    new_timeline.activate();
                }
            }
        }
@@ -267,10 +264,7 @@ impl UninitializedTimeline<'_> {
            .await
            .context("Failed to flush after basebackup import")?;

-        // Initialize without loading the layer map. We started with an empty layer map, and already
-        // updated it for the layers that we created during the import.
-        let mut timelines = self.owning_tenant.timelines.lock().unwrap();
-        self.initialize_with_lock(ctx, &mut timelines, false, true)
+        self.initialize(ctx)
    }

    fn raw_timeline(&self) -> anyhow::Result<&Arc<Timeline>> {
@@ -475,7 +469,7 @@ impl Tenant {
        local_metadata: Option<TimelineMetadata>,
        ancestor: Option<Arc<Timeline>>,
        first_save: bool,
-        ctx: &RequestContext,
+        _ctx: &RequestContext,
    ) -> anyhow::Result<()> {
        let tenant_id = self.tenant_id;

@@ -510,7 +504,7 @@ impl Tenant {
            // Do not start walreceiver here. We do need loaded layer map for reconcile_with_remote
            // But we shouldnt start walreceiver before we have all the data locally, because working walreceiver
            // will ingest data which may require looking at the layers which are not yet available locally
-            match timeline.initialize_with_lock(ctx, &mut timelines_accessor, true, false) {
+            match timeline.initialize_with_lock(&mut timelines_accessor, true, false) {
                Ok(new_timeline) => new_timeline,
                Err(e) => {
                    error!("Failed to initialize timeline {tenant_id}/{timeline_id}: {e:?}");
@@ -622,7 +616,7 @@ impl Tenant {
                match tenant_clone.attach(ctx).await {
                    Ok(_) => {}
                    Err(e) => {
-                        tenant_clone.set_broken(e.to_string());
+                        tenant_clone.set_broken(&e.to_string());
                        error!("error attaching tenant: {:?}", e);
                    }
                }
@@ -635,7 +629,7 @@ impl Tenant {
    ///
    /// Background task that downloads all data for a tenant and brings it to Active state.
    ///
-    #[instrument(skip_all, fields(tenant_id=%self.tenant_id))]
+    #[instrument(skip(self, ctx), fields(tenant_id=%self.tenant_id))]
    async fn attach(self: &Arc<Tenant>, ctx: RequestContext) -> anyhow::Result<()> {
        // Create directory with marker file to indicate attaching state.
        // The load_local_tenants() function in tenant::mgr relies on the marker file
@@ -756,7 +750,7 @@ impl Tenant {

        // Start background operations and open the tenant for business.
        // The loops will shut themselves down when they notice that the tenant is inactive.
-        self.activate(&ctx)?;
+        self.activate()?;

        info!("Done");

@@ -830,10 +824,7 @@ impl Tenant {
    pub fn create_broken_tenant(conf: &'static PageServerConf, tenant_id: TenantId) -> Arc<Tenant> {
        let wal_redo_manager = Arc::new(PostgresRedoManager::new(conf, tenant_id));
        Arc::new(Tenant::new(
-            TenantState::Broken {
-                reason: "create_broken_tenant".into(),
-                backtrace: String::new(),
-            },
+            TenantState::Broken,
            conf,
            TenantConfOpt::default(),
            wal_redo_manager,
@@ -894,7 +885,7 @@ impl Tenant {
                match tenant_clone.load(&ctx).await {
                    Ok(()) => {}
                    Err(err) => {
-                        tenant_clone.set_broken(err.to_string());
+                        tenant_clone.set_broken(&err.to_string());
                        error!("could not load tenant {tenant_id}: {err:?}");
                    }
                }
@@ -1031,7 +1022,7 @@ impl Tenant {

        // Start background operations and open the tenant for business.
        // The loops will shut themselves down when they notice that the tenant is inactive.
-        self.activate(ctx)?;
+        self.activate()?;

        info!("Done");

@@ -1367,7 +1358,12 @@ impl Tenant {

        // Stop the walreceiver first.
        debug!("waiting for wal receiver to shutdown");
-        timeline.walreceiver.stop().await;
+        task_mgr::shutdown_tasks(
+            Some(TaskKind::WalReceiverManager),
+            Some(self.tenant_id),
+            Some(timeline_id),
+        )
+        .await;
        debug!("wal receiver shutdown confirmed");

        info!("waiting for timeline tasks to shutdown");
@@ -1446,7 +1442,7 @@ impl Tenant {
    }

    pub fn current_state(&self) -> TenantState {
-        self.state.borrow().clone()
+        *self.state.borrow()
    }

    pub fn is_active(&self) -> bool {
@@ -1454,18 +1450,18 @@ impl Tenant {
    }

    /// Changes tenant status to active, unless shutdown was already requested.
-    fn activate(&self, ctx: &RequestContext) -> anyhow::Result<()> {
+    fn activate(&self) -> anyhow::Result<()> {
        let mut result = Ok(());
        self.state.send_modify(|current_state| {
-            match &*current_state {
+            match *current_state {
                TenantState::Active => {
                    // activate() was called on an already Active tenant. Shouldn't happen.
                    result = Err(anyhow::anyhow!("Tenant is already active"));
                }
-                TenantState::Broken { reason, .. } => {
+                TenantState::Broken => {
                    // This shouldn't happen either
                    result = Err(anyhow::anyhow!(
-                        "Could not activate tenant because it is in broken state due to: {reason}",
+                        "Could not activate tenant because it is in broken state"
                    ));
                }
                TenantState::Stopping => {
@@ -1488,23 +1484,7 @@ impl Tenant {
                    tasks::start_background_loops(self.tenant_id);

                    for timeline in not_broken_timelines {
-                        match timeline
-                            .activate(ctx)
-                            .context("timeline activation for activating tenant")
-                        {
-                            Ok(()) => {}
-                            Err(e) => {
-                                error!(
-                                    "Failed to activate timeline {}: {:#}",
-                                    timeline.timeline_id, e
-                                );
-                                timeline.set_state(TimelineState::Broken);
-                                *current_state = TenantState::broken_from_reason(format!(
-                                    "failed to activate timeline {}: {}",
-                                    timeline.timeline_id, e
-                                ));
-                            }
-                        }
+                        timeline.activate();
                    }
                }
            }
@@ -1515,7 +1495,7 @@ impl Tenant {
    /// Change tenant status to Stopping, to mark that it is being shut down
    pub fn set_stopping(&self) {
        self.state.send_modify(|current_state| {
-            match current_state {
+            match *current_state {
                TenantState::Active | TenantState::Loading | TenantState::Attaching => {
                    *current_state = TenantState::Stopping;

@@ -1531,8 +1511,8 @@ impl Tenant {
                        timeline.set_state(TimelineState::Stopping);
                    }
                }
-                TenantState::Broken { reason, .. } => {
-                    info!("Cannot set tenant to Stopping state, it is in Broken state due to: {reason}");
+                TenantState::Broken => {
+                    info!("Cannot set tenant to Stopping state, it is already in Broken state");
                }
                TenantState::Stopping => {
                    // The tenant was detached, or system shutdown was requested, while we were
@@ -1543,7 +1523,7 @@ impl Tenant {
        });
    }

-    pub fn set_broken(&self, reason: String) {
+    pub fn set_broken(&self, reason: &str) {
        self.state.send_modify(|current_state| {
            match *current_state {
                TenantState::Active => {
@@ -1551,24 +1531,24 @@ impl Tenant {
                    // while loading or attaching a tenant. A tenant that has already been
                    // activated should never be marked as broken. We cope with it the best
                    // we can, but it shouldn't happen.
+                    *current_state = TenantState::Broken;
                    warn!("Changing Active tenant to Broken state, reason: {}", reason);
-                    *current_state = TenantState::broken_from_reason(reason);
                }
-                TenantState::Broken { .. } => {
+                TenantState::Broken => {
                    // This shouldn't happen either
                    warn!("Tenant is already in Broken state");
                }
                TenantState::Stopping => {
                    // This shouldn't happen either
+                    *current_state = TenantState::Broken;
                    warn!(
                        "Marking Stopping tenant as Broken state, reason: {}",
                        reason
                    );
-                    *current_state = TenantState::broken_from_reason(reason);
                }
                TenantState::Loading | TenantState::Attaching => {
                    info!("Setting tenant as Broken state, reason: {}", reason);
-                    *current_state = TenantState::broken_from_reason(reason);
+                    *current_state = TenantState::Broken;
                }
            }
        });
@@ -1581,7 +1561,7 @@ impl Tenant {
    pub async fn wait_to_become_active(&self) -> anyhow::Result<()> {
        let mut receiver = self.state.subscribe();
        loop {
-            let current_state = receiver.borrow_and_update().clone();
+            let current_state = *receiver.borrow_and_update();
            match current_state {
                TenantState::Loading | TenantState::Attaching => {
                    // in these states, there's a chance that we can reach ::Active
@@ -1590,12 +1570,12 @@ impl Tenant {
                TenantState::Active { .. } => {
                    return Ok(());
                }
-                TenantState::Broken { .. } | TenantState::Stopping => {
+                TenantState::Broken | TenantState::Stopping => {
                    // There's no chance the tenant can transition back into ::Active
                    anyhow::bail!(
                        "Tenant {} will not become active. Current state: {:?}",
                        self.tenant_id,
-                        &current_state,
+                        current_state,
                    );
                }
            }
@@ -1726,13 +1706,6 @@ impl Tenant {
            .unwrap_or(self.conf.default_tenant_conf.trace_read_requests)
    }

-    pub fn get_min_resident_size_override(&self) -> Option<u64> {
-        let tenant_conf = self.tenant_conf.read().unwrap();
-        tenant_conf
-            .min_resident_size_override
-            .or(self.conf.default_tenant_conf.min_resident_size_override)
-    }
-
    pub fn set_new_tenant_config(&self, new_tenant_conf: TenantConfOpt) {
        *self.tenant_conf.write().unwrap() = new_tenant_conf;
    }
@@ -1776,23 +1749,21 @@ impl Tenant {
        let (state, mut rx) = watch::channel(state);

        tokio::spawn(async move {
-            let mut current_state: &'static str = From::from(&*rx.borrow_and_update());
+            let current_state = *rx.borrow_and_update();
            let tid = tenant_id.to_string();
            TENANT_STATE_METRIC
-                .with_label_values(&[&tid, current_state])
+                .with_label_values(&[&tid, current_state.as_str()])
                .inc();
            loop {
                match rx.changed().await {
                    Ok(()) => {
-                        let new_state: &'static str = From::from(&*rx.borrow_and_update());
+                        let new_state = *rx.borrow();
                        TENANT_STATE_METRIC
-                            .with_label_values(&[&tid, current_state])
+                            .with_label_values(&[&tid, current_state.as_str()])
                            .dec();
                        TENANT_STATE_METRIC
-                            .with_label_values(&[&tid, new_state])
+                            .with_label_values(&[&tid, new_state.as_str()])
                            .inc();
-
-                        current_state = new_state;
                    }
                    Err(_sender_dropped_error) => {
                        info!("Tenant dropped the state updates sender, quitting waiting for tenant state change");
@@ -2115,7 +2086,7 @@ impl Tenant {
        src_timeline: &Arc<Timeline>,
        dst_id: TimelineId,
        start_lsn: Option<Lsn>,
-        ctx: &RequestContext,
+        _ctx: &RequestContext,
    ) -> anyhow::Result<Arc<Timeline>> {
        let src_id = src_timeline.timeline_id;

@@ -2208,7 +2179,7 @@ impl Tenant {
                false,
                Some(Arc::clone(src_timeline)),
            )?
-            .initialize_with_lock(ctx, &mut timelines, true, true)?;
+            .initialize_with_lock(&mut timelines, true, true)?;
        drop(timelines);

        // Root timeline gets its layers during creation and uploads them along with the metadata.
@@ -2319,11 +2290,9 @@ impl Tenant {
                )
            })?;

-        // Initialize the timeline without loading the layer map, because we already updated the layer
-        // map above, when we imported the datadir.
        let timeline = {
            let mut timelines = self.timelines.lock().unwrap();
-            raw_timeline.initialize_with_lock(ctx, &mut timelines, false, true)?
+            raw_timeline.initialize_with_lock(&mut timelines, false, true)?
        };

        info!(
@@ -2814,7 +2783,6 @@ pub mod harness {
                max_lsn_wal_lag: Some(tenant_conf.max_lsn_wal_lag),
                trace_read_requests: Some(tenant_conf.trace_read_requests),
                eviction_policy: Some(tenant_conf.eviction_policy),
-                min_resident_size_override: tenant_conf.min_resident_size_override,
            }
        }
    }
--- a/pageserver/src/tenant/config.rs
+++ b/pageserver/src/tenant/config.rs
@@ -92,7 +92,6 @@ pub struct TenantConf {
    pub max_lsn_wal_lag: NonZeroU64,
    pub trace_read_requests: bool,
    pub eviction_policy: EvictionPolicy,
-    pub min_resident_size_override: Option<u64>,
 }

 /// Same as TenantConf, but this struct preserves the information about
@@ -160,10 +159,6 @@ pub struct TenantConfOpt {
    #[serde(skip_serializing_if = "Option::is_none")]
    #[serde(default)]
    pub eviction_policy: Option<EvictionPolicy>,
-
-    #[serde(skip_serializing_if = "Option::is_none")]
-    #[serde(default)]
-    pub min_resident_size_override: Option<u64>,
 }

 #[derive(Debug, Clone, Copy, PartialEq, Eq, Serialize, Deserialize)]
@@ -225,9 +220,6 @@ impl TenantConfOpt {
                .trace_read_requests
                .unwrap_or(global_conf.trace_read_requests),
            eviction_policy: self.eviction_policy.unwrap_or(global_conf.eviction_policy),
-            min_resident_size_override: self
-                .min_resident_size_override
-                .or(global_conf.min_resident_size_override),
        }
    }
 }
@@ -259,7 +251,6 @@ impl Default for TenantConf {
                .expect("cannot parse default max walreceiver Lsn wal lag"),
            trace_read_requests: false,
            eviction_policy: EvictionPolicy::NoEviction,
-            min_resident_size_override: None,
        }
    }
 }
--- a/pageserver/src/tenant/layer_map.rs
+++ b/pageserver/src/tenant/layer_map.rs
@@ -52,7 +52,7 @@ use crate::metrics::NUM_ONDISK_LAYERS;
 use crate::repository::Key;
 use crate::tenant::storage_layer::InMemoryLayer;
 use crate::tenant::storage_layer::Layer;
-use anyhow::{bail, Result};
+use anyhow::Result;
 use std::collections::VecDeque;
 use std::ops::Range;
 use std::sync::Arc;
@@ -126,7 +126,7 @@ where
    ///
    /// Insert an on-disk layer.
    ///
-    pub fn insert_historic(&mut self, layer: Arc<L>) -> anyhow::Result<()> {
+    pub fn insert_historic(&mut self, layer: Arc<L>) {
        self.layer_map.insert_historic_noflush(layer)
    }

@@ -274,22 +274,17 @@ where
    ///
    /// Helper function for BatchedUpdates::insert_historic
    ///
-    pub(self) fn insert_historic_noflush(&mut self, layer: Arc<L>) -> anyhow::Result<()> {
-        let key = historic_layer_coverage::LayerKey::from(&*layer);
-        if self.historic.contains(&key) {
-            bail!(
-                "Attempt to insert duplicate layer {} in layer map",
-                layer.short_id()
-            );
-        }
-        self.historic.insert(key, Arc::clone(&layer));
+    pub(self) fn insert_historic_noflush(&mut self, layer: Arc<L>) {
+        self.historic.insert(
+            historic_layer_coverage::LayerKey::from(&*layer),
+            Arc::clone(&layer),
+        );

        if Self::is_l0(&layer) {
            self.l0_delta_layers.push(layer);
        }

        NUM_ONDISK_LAYERS.inc();
-        Ok(())
    }

    ///
@@ -843,7 +838,7 @@ mod tests {

            let expected_in_counts = (1, usize::from(expected_l0));

-            map.batch_update().insert_historic(remote.clone()).unwrap();
+            map.batch_update().insert_historic(remote.clone());
            assert_eq!(count_layer_in(&map, &remote), expected_in_counts);

            let replaced = map
--- a/pageserver/src/tenant/layer_map/historic_layer_coverage.rs
+++ b/pageserver/src/tenant/layer_map/historic_layer_coverage.rs
@@ -417,14 +417,6 @@ impl<Value: Clone> BufferedHistoricLayerCoverage<Value> {
        }
    }

-    pub fn contains(&self, layer_key: &LayerKey) -> bool {
-        match self.buffer.get(layer_key) {
-            Some(None) => false,                         // layer remove was buffered
-            Some(_) => true,                             // layer insert was buffered
-            None => self.layers.contains_key(layer_key), // no buffered ops for this layer
-        }
-    }
-
    pub fn insert(&mut self, layer_key: LayerKey, value: Value) {
        self.buffer.insert(layer_key, Some(value));
    }
--- a/pageserver/src/tenant/mgr.rs
+++ b/pageserver/src/tenant/mgr.rs
@@ -537,7 +537,7 @@ where
            Some(tenant) => match tenant.current_state() {
                TenantState::Attaching
                | TenantState::Loading
-                | TenantState::Broken { .. }
+                | TenantState::Broken
                | TenantState::Active => tenant.set_stopping(),
                TenantState::Stopping => return Err(TenantStateError::IsStopping(tenant_id)),
            },
@@ -565,7 +565,7 @@ where
            let tenants_accessor = TENANTS.read().await;
            match tenants_accessor.get(&tenant_id) {
                Some(tenant) => {
-                    tenant.set_broken(e.to_string());
+                    tenant.set_broken(&e.to_string());
                }
                None => {
                    warn!("Tenant {tenant_id} got removed from memory");
--- a/pageserver/src/tenant/storage_layer.rs
+++ b/pageserver/src/tenant/storage_layer.rs
@@ -121,10 +121,10 @@ struct LayerAccessStatsInner {
 }

 #[derive(Debug, Clone, Copy)]
-pub(crate) struct LayerAccessStatFullDetails {
-    pub(crate) when: SystemTime,
-    pub(crate) task_kind: TaskKind,
-    pub(crate) access_kind: LayerAccessKind,
+pub(super) struct LayerAccessStatFullDetails {
+    pub(super) when: SystemTime,
+    pub(super) task_kind: TaskKind,
+    pub(super) access_kind: LayerAccessKind,
 }

 #[derive(Clone, Copy, strum_macros::EnumString)]
@@ -255,7 +255,7 @@ impl LayerAccessStats {
        ret
    }

-    fn most_recent_access_or_residence_event(
+    pub(super) fn most_recent_access_or_residence_event(
        &self,
    ) -> Either<LayerAccessStatFullDetails, LayerResidenceEvent> {
        let locked = self.0.lock().unwrap();
@@ -268,13 +268,6 @@ impl LayerAccessStats {
            }
        }
    }
-
-    pub(crate) fn latest_activity(&self) -> SystemTime {
-        match self.most_recent_access_or_residence_event() {
-            Either::Left(mra) => mra.when,
-            Either::Right(re) => re.timestamp,
-        }
-    }
 }

 /// Supertrait of the [`Layer`] trait that captures the bare minimum interface
--- a/pageserver/src/tenant/tasks.rs
+++ b/pageserver/src/tenant/tasks.rs
@@ -209,7 +209,7 @@ async fn wait_for_active_tenant(
        loop {
            match tenant_state_updates.changed().await {
                Ok(()) => {
-                    let new_state = &*tenant_state_updates.borrow();
+                    let new_state = *tenant_state_updates.borrow();
                    match new_state {
                        TenantState::Active => {
                            debug!("Tenant state changed to active, continuing the task loop");
--- a/pageserver/src/tenant/timeline.rs
+++ b/pageserver/src/tenant/timeline.rs
@@ -13,8 +13,6 @@ use pageserver_api::models::{
    DownloadRemoteLayersTaskInfo, DownloadRemoteLayersTaskSpawnRequest,
    DownloadRemoteLayersTaskState, LayerMapInfo, LayerResidenceStatus, TimelineState,
 };
-use remote_storage::GenericRemoteStorage;
-use storage_broker::BrokerClientChannel;
 use tokio::sync::{oneshot, watch, Semaphore, TryAcquireError};
 use tokio_util::sync::CancellationToken;
 use tracing::*;
@@ -31,7 +29,7 @@ use std::sync::atomic::{AtomicI64, Ordering as AtomicOrdering};
 use std::sync::{Arc, Mutex, MutexGuard, RwLock, Weak};
 use std::time::{Duration, Instant, SystemTime};

-use crate::broker_client::{get_broker_client, is_broker_client_initialized};
+use crate::broker_client::is_broker_client_initialized;
 use crate::context::{DownloadBehavior, RequestContext};
 use crate::tenant::remote_timeline_client::{self, index::LayerFileMetadata};
 use crate::tenant::storage_layer::{
@@ -72,10 +70,10 @@ use crate::walredo::WalRedoManager;
 use crate::METADATA_FILE_NAME;
 use crate::ZERO_PAGE;
 use crate::{is_temporary, task_mgr};
+use walreceiver::spawn_connection_manager_task;

 pub(super) use self::eviction_task::EvictionTaskTenantState;
 use self::eviction_task::EvictionTaskTimelineState;
-use self::walreceiver::{WalReceiver, WalReceiverConf};

 use super::layer_map::BatchedUpdates;
 use super::remote_timeline_client::index::IndexPart;
@@ -145,7 +143,7 @@ pub struct Timeline {
    // 'last_record_lsn.load().prev'. It's used to set the xl_prev pointer of the
    // first WAL record when the node is started up. But here, we just
    // keep track of it.
-    last_record_lsn: SeqWait<RecordLsn, Lsn, ()>,
+    last_record_lsn: SeqWait<RecordLsn, Lsn>,

    // All WAL records have been processed and stored durably on files on
    // local disk, up to this LSN. On crash and restart, we need to re-process
@@ -215,7 +213,6 @@ pub struct Timeline {
    /// or None if WAL receiver has not received anything for this timeline
    /// yet.
    pub last_received_wal: Mutex<Option<WalReceiverInfo>>,
-    pub walreceiver: WalReceiver,

    /// Relation size cache
    pub rel_size_cache: RwLock<HashMap<RelTag, (Lsn, BlockNumber)>>,
@@ -868,18 +865,10 @@ impl Timeline {
        Ok(())
    }

-    pub fn activate(self: &Arc<Self>, ctx: &RequestContext) -> anyhow::Result<()> {
-        if is_broker_client_initialized() {
-            self.launch_wal_receiver(ctx, get_broker_client().clone())?;
-        } else if cfg!(test) {
-            info!("not launching WAL receiver because broker client hasn't been initialized");
-        } else {
-            anyhow::bail!("broker client not initialized");
-        }
-
+    pub fn activate(self: &Arc<Self>) {
        self.set_state(TimelineState::Active);
+        self.launch_wal_receiver();
        self.launch_eviction_task();
-        Ok(())
    }

    pub fn set_state(&self, new_state: TimelineState) {
@@ -968,25 +957,6 @@ impl Timeline {
        }
    }

-    /// Evict a batch of layers.
-    ///
-    /// GenericRemoteStorage reference is required as a witness[^witness_article] for "remote storage is configured."
-    ///
-    /// [^witness_article]: https://willcrichton.net/rust-api-type-patterns/witnesses.html
-    pub async fn evict_layers(
-        &self,
-        _: &GenericRemoteStorage,
-        layers_to_evict: &[Arc<dyn PersistentLayer>],
-        cancel: CancellationToken,
-    ) -> anyhow::Result<Vec<Option<anyhow::Result<bool>>>> {
-        let remote_client = self.remote_client.clone().expect(
-            "GenericRemoteStorage is configured, so timeline must have RemoteTimelineClient",
-        );
-
-        self.evict_layer_batch(&remote_client, layers_to_evict, cancel)
-            .await
-    }
-
    /// Evict multiple layers at once, continuing through errors.
    ///
    /// Try to evict the given `layers_to_evict` by
@@ -1024,15 +994,6 @@ impl Timeline {
        // now lock out layer removal (compaction, gc, timeline deletion)
        let layer_removal_guard = self.layer_removal_cs.lock().await;

-        {
-            // to avoid racing with detach and delete_timeline
-            let state = self.current_state();
-            anyhow::ensure!(
-                state == TimelineState::Active,
-                "timeline is not active but {state:?}"
-            );
-        }
-
        // start the batch update
        let mut layer_map = self.layers.write().unwrap();
        let mut batch_updates = layer_map.batch_update();
@@ -1066,8 +1027,6 @@ impl Timeline {
        use super::layer_map::Replacement;

        if local_layer.is_remote_layer() {
-            // TODO(issue #3851): consider returning an err here instead of false,
-            // which is the same out the match later
            return Ok(false);
        }

@@ -1137,9 +1096,6 @@ impl Timeline {
                    self.metrics
                        .evictions_with_low_residence_duration
                        .observe(delta);
-                    info!(layer=%local_layer.short_id(), residence_millis=delta.as_millis(), "evicted layer after known residence period");
-                } else {
-                    info!(layer=%local_layer.short_id(), "evicted layer after unknown residence period");
                }

                true
@@ -1230,31 +1186,7 @@ impl Timeline {
        let (layer_flush_start_tx, _) = tokio::sync::watch::channel(0);
        let (layer_flush_done_tx, _) = tokio::sync::watch::channel((0, Ok(())));

-        let tenant_conf_guard = tenant_conf.read().unwrap();
-        let wal_connect_timeout = tenant_conf_guard
-            .walreceiver_connect_timeout
-            .unwrap_or(conf.default_tenant_conf.walreceiver_connect_timeout);
-        let lagging_wal_timeout = tenant_conf_guard
-            .lagging_wal_timeout
-            .unwrap_or(conf.default_tenant_conf.lagging_wal_timeout);
-        let max_lsn_wal_lag = tenant_conf_guard
-            .max_lsn_wal_lag
-            .unwrap_or(conf.default_tenant_conf.max_lsn_wal_lag);
-        drop(tenant_conf_guard);
-
        Arc::new_cyclic(|myself| {
-            let walreceiver = WalReceiver::new(
-                TenantTimelineId::new(tenant_id, timeline_id),
-                Weak::clone(myself),
-                WalReceiverConf {
-                    wal_connect_timeout,
-                    lagging_wal_timeout,
-                    max_lsn_wal_lag,
-                    auth_token: crate::config::SAFEKEEPER_AUTH_TOKEN.get().cloned(),
-                    availability_zone: conf.availability_zone.clone(),
-                },
-            );
-
            let mut result = Timeline {
                conf,
                tenant_conf,
@@ -1265,18 +1197,14 @@ impl Timeline {
                layers: RwLock::new(LayerMap::default()),

                walredo_mgr,
-                walreceiver,

                remote_client: remote_client.map(Arc::new),

                // initialize in-memory 'last_record_lsn' from 'disk_consistent_lsn'.
-                last_record_lsn: SeqWait::new(
-                    RecordLsn {
-                        last: disk_consistent_lsn,
-                        prev: metadata.prev_record_lsn().unwrap_or(Lsn(0)),
-                    },
-                    (),
-                ),
+                last_record_lsn: SeqWait::new(RecordLsn {
+                    last: disk_consistent_lsn,
+                    prev: metadata.prev_record_lsn().unwrap_or(Lsn(0)),
+                }),
                disk_consistent_lsn: AtomicLsn::new(disk_consistent_lsn.0),

                last_freeze_at: AtomicLsn::new(disk_consistent_lsn.0),
@@ -1388,17 +1316,44 @@ impl Timeline {
        *flush_loop_state = FlushLoopState::Running;
    }

-    pub(super) fn launch_wal_receiver(
-        &self,
-        ctx: &RequestContext,
-        broker_client: BrokerClientChannel,
-    ) -> anyhow::Result<()> {
+    pub(super) fn launch_wal_receiver(self: &Arc<Self>) {
+        if !is_broker_client_initialized() {
+            if cfg!(test) {
+                info!("not launching WAL receiver because broker client hasn't been initialized");
+                return;
+            } else {
+                panic!("broker client not initialized");
+            }
+        }
+
        info!(
            "launching WAL receiver for timeline {} of tenant {}",
            self.timeline_id, self.tenant_id
        );
-        self.walreceiver.start(ctx, broker_client)?;
-        Ok(())
+        let tenant_conf_guard = self.tenant_conf.read().unwrap();
+        let lagging_wal_timeout = tenant_conf_guard
+            .lagging_wal_timeout
+            .unwrap_or(self.conf.default_tenant_conf.lagging_wal_timeout);
+        let walreceiver_connect_timeout = tenant_conf_guard
+            .walreceiver_connect_timeout
+            .unwrap_or(self.conf.default_tenant_conf.walreceiver_connect_timeout);
+        let max_lsn_wal_lag = tenant_conf_guard
+            .max_lsn_wal_lag
+            .unwrap_or(self.conf.default_tenant_conf.max_lsn_wal_lag);
+        drop(tenant_conf_guard);
+        let self_clone = Arc::clone(self);
+        let background_ctx =
+            // XXX: this is a detached_child. Plumb through the ctx from call sites.
+            RequestContext::todo_child(TaskKind::WalReceiverManager, DownloadBehavior::Error);
+        spawn_connection_manager_task(
+            self_clone,
+            walreceiver_connect_timeout,
+            lagging_wal_timeout,
+            max_lsn_wal_lag,
+            crate::config::SAFEKEEPER_AUTH_TOKEN.get().cloned(),
+            self.conf.availability_zone.clone(),
+            background_ctx,
+        );
    }

    ///
@@ -1449,7 +1404,7 @@ impl Timeline {

                trace!("found layer {}", layer.path().display());
                total_physical_size += file_size;
-                updates.insert_historic(Arc::new(layer))?;
+                updates.insert_historic(Arc::new(layer));
                num_layers += 1;
            } else if let Some(deltafilename) = DeltaFileName::parse_str(&fname) {
                // Create a DeltaLayer struct for each delta file.
@@ -1481,7 +1436,7 @@ impl Timeline {

                trace!("found layer {}", layer.path().display());
                total_physical_size += file_size;
-                updates.insert_historic(Arc::new(layer))?;
+                updates.insert_historic(Arc::new(layer));
                num_layers += 1;
            } else if fname == METADATA_FILE_NAME || fname.ends_with(".old") {
                // ignore these
@@ -1555,7 +1510,7 @@ impl Timeline {
            // remote index file?
            // If so, rename_to_backup those files & replace their local layer with
            // a RemoteLayer in the layer map so that we re-download them on-demand.
-            if let Some(local_layer) = &local_layer {
+            if let Some(local_layer) = local_layer {
                let local_layer_path = local_layer
                    .local_path()
                    .expect("caller must ensure that local_layers only contains local layers");
@@ -1580,6 +1535,7 @@ impl Timeline {
                        anyhow::bail!("could not rename file {local_layer_path:?}: {err:?}");
                    } else {
                        self.metrics.resident_physical_size_gauge.sub(local_size);
+                        updates.remove_historic(local_layer);
                        // fall-through to adding the remote layer
                    }
                } else {
@@ -1615,11 +1571,7 @@ impl Timeline {
                    );
                    let remote_layer = Arc::new(remote_layer);

-                    if let Some(local_layer) = &local_layer {
-                        updates.replace_historic(local_layer, remote_layer)?;
-                    } else {
-                        updates.insert_historic(remote_layer)?;
-                    }
+                    updates.insert_historic(remote_layer);
                }
                LayerFileName::Delta(deltafilename) => {
                    // Create a RemoteLayer for the delta file.
@@ -1643,11 +1595,7 @@ impl Timeline {
                        LayerAccessStats::for_loading_layer(LayerResidenceStatus::Evicted),
                    );
                    let remote_layer = Arc::new(remote_layer);
-                    if let Some(local_layer) = &local_layer {
-                        updates.replace_historic(local_layer, remote_layer)?;
-                    } else {
-                        updates.insert_historic(remote_layer)?;
-                    }
+                    updates.insert_historic(remote_layer);
                }
            }
        }
@@ -2423,7 +2371,7 @@ impl Timeline {
        assert!(new_lsn.is_aligned());

        self.metrics.last_record_gauge.set(new_lsn.0 as i64);
-        self.last_record_lsn.advance(new_lsn, None);
+        self.last_record_lsn.advance(new_lsn);
    }

    fn freeze_inmem_layer(&self, write_lock_held: bool) {
@@ -2694,7 +2642,7 @@ impl Timeline {
            .write()
            .unwrap()
            .batch_update()
-            .insert_historic(Arc::new(new_delta))?;
+            .insert_historic(Arc::new(new_delta));

        // update the timeline's physical size
        let sz = new_delta_path.metadata()?.len();
@@ -2899,7 +2847,7 @@ impl Timeline {
            self.metrics
                .resident_physical_size_gauge
                .add(metadata.len());
-            updates.insert_historic(Arc::new(l))?;
+            updates.insert_historic(Arc::new(l));
        }
        updates.flush();
        drop(layers);
@@ -3332,7 +3280,7 @@ impl Timeline {

            new_layer_paths.insert(new_delta_path, LayerFileMetadata::new(metadata.len()));
            let x: Arc<dyn PersistentLayer + 'static> = Arc::new(l);
-            updates.insert_historic(x)?;
+            updates.insert_historic(x);
        }

        // Now that we have reshuffled the data to set of new delta layers, we can
@@ -4064,67 +4012,6 @@ impl Timeline {
    }
 }

-pub struct DiskUsageEvictionInfo {
-    /// Timeline's largest layer (remote or resident)
-    pub max_layer_size: Option<u64>,
-    /// Timeline's resident layers
-    pub resident_layers: Vec<LocalLayerInfoForDiskUsageEviction>,
-}
-
-pub struct LocalLayerInfoForDiskUsageEviction {
-    pub layer: Arc<dyn PersistentLayer>,
-    pub last_activity_ts: SystemTime,
-}
-
-impl std::fmt::Debug for LocalLayerInfoForDiskUsageEviction {
-    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
-        // format the tv_sec, tv_nsec into rfc3339 in case someone is looking at it
-        // having to allocate a string to this is bad, but it will rarely be formatted
-        let ts = chrono::DateTime::<chrono::Utc>::from(self.last_activity_ts);
-        let ts = ts.to_rfc3339_opts(chrono::SecondsFormat::Nanos, true);
-        f.debug_struct("LocalLayerInfoForDiskUsageEviction")
-            .field("layer", &self.layer)
-            .field("last_activity", &ts)
-            .finish()
-    }
-}
-
-impl LocalLayerInfoForDiskUsageEviction {
-    pub fn file_size(&self) -> u64 {
-        self.layer.file_size()
-    }
-}
-
-impl Timeline {
-    pub(crate) fn get_local_layers_for_disk_usage_eviction(&self) -> DiskUsageEvictionInfo {
-        let layers = self.layers.read().unwrap();
-
-        let mut max_layer_size: Option<u64> = None;
-        let mut resident_layers = Vec::new();
-
-        for l in layers.iter_historic_layers() {
-            let file_size = l.file_size();
-            max_layer_size = max_layer_size.map_or(Some(file_size), |m| Some(m.max(file_size)));
-
-            if l.is_remote_layer() {
-                continue;
-            }
-
-            let last_activity_ts = l.access_stats().latest_activity();
-
-            resident_layers.push(LocalLayerInfoForDiskUsageEviction {
-                layer: l,
-                last_activity_ts,
-            });
-        }
-
-        DiskUsageEvictionInfo {
-            max_layer_size,
-            resident_layers,
-        }
-    }
-}
-
 type TraversalPathItem = (
    ValueReconstructResult,
    Lsn,
--- a/pageserver/src/tenant/timeline/eviction_task.rs
+++ b/pageserver/src/tenant/timeline/eviction_task.rs
@@ -20,6 +20,7 @@ use std::{
    time::{Duration, SystemTime},
 };

+use either::Either;
 use tokio::time::Instant;
 use tokio_util::sync::CancellationToken;
 use tracing::{debug, error, info, instrument, warn};
@@ -184,7 +185,13 @@ impl Timeline {
                if hist_layer.is_remote_layer() {
                    continue;
                }
-                let last_activity_ts = hist_layer.access_stats().latest_activity();
+                let last_activity_ts = match hist_layer
+                    .access_stats()
+                    .most_recent_access_or_residence_event()
+                {
+                    Either::Left(mra) => mra.when,
+                    Either::Right(re) => re.timestamp,
+                };
                let no_activity_for = match now.duration_since(last_activity_ts) {
                    Ok(d) => d,
                    Err(_e) => {
--- a/pageserver/src/tenant/timeline/walreceiver.rs
+++ b/pageserver/src/tenant/timeline/walreceiver.rs
@@ -23,133 +23,14 @@
 mod connection_manager;
 mod walreceiver_connection;

-use crate::context::{DownloadBehavior, RequestContext};
-use crate::task_mgr::{self, TaskKind, WALRECEIVER_RUNTIME};
-use crate::tenant::timeline::walreceiver::connection_manager::{
-    connection_manager_loop_step, ConnectionManagerState,
-};
+use crate::task_mgr::WALRECEIVER_RUNTIME;

-use anyhow::Context;
 use std::future::Future;
-use std::num::NonZeroU64;
-use std::ops::ControlFlow;
-use std::sync::atomic::{self, AtomicBool};
-use std::sync::{Arc, Weak};
-use std::time::Duration;
-use storage_broker::BrokerClientChannel;
-use tokio::select;
 use tokio::sync::watch;
 use tokio_util::sync::CancellationToken;
 use tracing::*;

-use utils::id::TenantTimelineId;
-
-use super::Timeline;
-
-#[derive(Clone)]
-pub struct WalReceiverConf {
-    /// The timeout on the connection to safekeeper for WAL streaming.
-    pub wal_connect_timeout: Duration,
-    /// The timeout to use to determine when the current connection is "stale" and reconnect to the other one.
-    pub lagging_wal_timeout: Duration,
-    /// The Lsn lag to use to determine when the current connection is lagging to much behind and reconnect to the other one.
-    pub max_lsn_wal_lag: NonZeroU64,
-    pub auth_token: Option<Arc<String>>,
-    pub availability_zone: Option<String>,
-}
-
-pub struct WalReceiver {
-    timeline: TenantTimelineId,
-    timeline_ref: Weak<Timeline>,
-    conf: WalReceiverConf,
-    started: AtomicBool,
-}
-
-impl WalReceiver {
-    pub fn new(
-        timeline: TenantTimelineId,
-        timeline_ref: Weak<Timeline>,
-        conf: WalReceiverConf,
-    ) -> Self {
-        Self {
-            timeline,
-            timeline_ref,
-            conf,
-            started: AtomicBool::new(false),
-        }
-    }
-
-    pub fn start(
-        &self,
-        ctx: &RequestContext,
-        mut broker_client: BrokerClientChannel,
-    ) -> anyhow::Result<()> {
-        if self.started.load(atomic::Ordering::Acquire) {
-            anyhow::bail!("Wal receiver is already started");
-        }
-
-        let timeline = self.timeline_ref.upgrade().with_context(|| {
-            format!("walreceiver start on a dropped timeline {}", self.timeline)
-        })?;
-
-        let tenant_id = timeline.tenant_id;
-        let timeline_id = timeline.timeline_id;
-        let walreceiver_ctx =
-            ctx.detached_child(TaskKind::WalReceiverManager, DownloadBehavior::Error);
-
-        let wal_receiver_conf = self.conf.clone();
-        task_mgr::spawn(
-            WALRECEIVER_RUNTIME.handle(),
-            TaskKind::WalReceiverManager,
-            Some(tenant_id),
-            Some(timeline_id),
-            &format!("walreceiver for timeline {tenant_id}/{timeline_id}"),
-            false,
-            async move {
-                info!("WAL receiver manager started, connecting to broker");
-                let mut connection_manager_state = ConnectionManagerState::new(
-                    timeline,
-                    wal_receiver_conf,
-                );
-                loop {
-                    select! {
-                        _ = task_mgr::shutdown_watcher() => {
-                            info!("WAL receiver shutdown requested, shutting down");
-                            connection_manager_state.shutdown().await;
-                            return Ok(());
-                        },
-                        loop_step_result = connection_manager_loop_step(
-                            &mut broker_client,
-                            &mut connection_manager_state,
-                            &walreceiver_ctx,
-                        ) => match loop_step_result {
-                            ControlFlow::Continue(()) => continue,
-                            ControlFlow::Break(()) => {
-                                info!("Connection manager loop ended, shutting down");
-                                connection_manager_state.shutdown().await;
-                                return Ok(());
-                            }
-                        },
-                    }
-                }
-            }.instrument(info_span!(parent: None, "wal_connection_manager", tenant = %tenant_id, timeline = %timeline_id))
-        );
-
-        self.started.store(true, atomic::Ordering::Release);
-
-        Ok(())
-    }
-
-    pub async fn stop(&self) {
-        task_mgr::shutdown_tasks(
-            Some(TaskKind::WalReceiverManager),
-            Some(self.timeline.tenant_id),
-            Some(self.timeline.timeline_id),
-        )
-        .await;
-        self.started.store(false, atomic::Ordering::Release);
-    }
-}
+pub use connection_manager::spawn_connection_manager_task;

 /// A handle of an asynchronous task.
 /// The task has a channel that it can use to communicate its lifecycle events in a certain form, see [`TaskEvent`]
@@ -158,26 +39,26 @@ impl WalReceiver {
 /// Note that the communication happens via the `watch` channel, that does not accumulate the events, replacing the old one with the never one on submission.
 /// That may lead to certain events not being observed by the listener.
 #[derive(Debug)]
-struct TaskHandle<E> {
+pub struct TaskHandle<E> {
    join_handle: Option<tokio::task::JoinHandle<anyhow::Result<()>>>,
    events_receiver: watch::Receiver<TaskStateUpdate<E>>,
    cancellation: CancellationToken,
 }

-enum TaskEvent<E> {
+pub enum TaskEvent<E> {
    Update(TaskStateUpdate<E>),
    End(anyhow::Result<()>),
 }

 #[derive(Debug, Clone)]
-enum TaskStateUpdate<E> {
+pub enum TaskStateUpdate<E> {
    Started,
    Progress(E),
 }

 impl<E: Clone> TaskHandle<E> {
    /// Initializes the task, starting it immediately after the creation.
-    fn spawn<Fut>(
+    pub fn spawn<Fut>(
        task: impl FnOnce(watch::Sender<TaskStateUpdate<E>>, CancellationToken) -> Fut + Send + 'static,
    ) -> Self
    where
@@ -250,7 +131,7 @@ impl<E: Clone> TaskHandle<E> {
    }

    /// Aborts current task, waiting for it to finish.
-    async fn shutdown(self) {
+    pub async fn shutdown(self) {
        if let Some(jh) = self.join_handle {
            self.cancellation.cancel();
            match jh.await {
--- a/pageserver/src/tenant/timeline/walreceiver/connection_manager.rs
+++ b/pageserver/src/tenant/timeline/walreceiver/connection_manager.rs
@@ -11,9 +11,11 @@

 use std::{collections::HashMap, num::NonZeroU64, ops::ControlFlow, sync::Arc, time::Duration};

-use super::{TaskStateUpdate, WalReceiverConf};
+use super::TaskStateUpdate;
+use crate::broker_client::get_broker_client;
 use crate::context::{DownloadBehavior, RequestContext};
-use crate::task_mgr::TaskKind;
+use crate::task_mgr::WALRECEIVER_RUNTIME;
+use crate::task_mgr::{self, TaskKind};
 use crate::tenant::Timeline;
 use anyhow::Context;
 use chrono::{NaiveDateTime, Utc};
@@ -36,17 +38,75 @@ use utils::{

 use super::{walreceiver_connection::WalConnectionStatus, TaskEvent, TaskHandle};

+/// Spawns the loop to take care of the timeline's WAL streaming connection.
+pub fn spawn_connection_manager_task(
+    timeline: Arc<Timeline>,
+    wal_connect_timeout: Duration,
+    lagging_wal_timeout: Duration,
+    max_lsn_wal_lag: NonZeroU64,
+    auth_token: Option<Arc<String>>,
+    availability_zone: Option<String>,
+    ctx: RequestContext,
+) {
+    let mut broker_client = get_broker_client().clone();
+
+    let tenant_id = timeline.tenant_id;
+    let timeline_id = timeline.timeline_id;
+
+    task_mgr::spawn(
+        WALRECEIVER_RUNTIME.handle(),
+        TaskKind::WalReceiverManager,
+        Some(tenant_id),
+        Some(timeline_id),
+        &format!("walreceiver for timeline {tenant_id}/{timeline_id}"),
+        false,
+        async move {
+            info!("WAL receiver manager started, connecting to broker");
+            let mut walreceiver_state = WalreceiverState::new(
+                timeline,
+                wal_connect_timeout,
+                lagging_wal_timeout,
+                max_lsn_wal_lag,
+                auth_token,
+                availability_zone,
+            );
+            loop {
+                select! {
+                    _ = task_mgr::shutdown_watcher() => {
+                        info!("WAL receiver shutdown requested, shutting down");
+                        walreceiver_state.shutdown().await;
+                        return Ok(());
+                    },
+                    loop_step_result = connection_manager_loop_step(
+                        &mut broker_client,
+                        &mut walreceiver_state,
+                        &ctx,
+                    ) => match loop_step_result {
+                        ControlFlow::Continue(()) => continue,
+                        ControlFlow::Break(()) => {
+                            info!("Connection manager loop ended, shutting down");
+                            walreceiver_state.shutdown().await;
+                            return Ok(());
+                        }
+                    },
+                }
+            }
+        }
+        .instrument(
+            info_span!(parent: None, "wal_connection_manager", tenant = %tenant_id, timeline = %timeline_id),
+        ),
+    );
+}
+
 /// Attempts to subscribe for timeline updates, pushed by safekeepers into the broker.
 /// Based on the updates, desides whether to start, keep or stop a WAL receiver task.
 /// If storage broker subscription is cancelled, exits.
-pub(super) async fn connection_manager_loop_step(
+async fn connection_manager_loop_step(
    broker_client: &mut BrokerClientChannel,
-    connection_manager_state: &mut ConnectionManagerState,
+    walreceiver_state: &mut WalreceiverState,
    ctx: &RequestContext,
 ) -> ControlFlow<(), ()> {
-    let mut timeline_state_updates = connection_manager_state
-        .timeline
-        .subscribe_for_state_updates();
+    let mut timeline_state_updates = walreceiver_state.timeline.subscribe_for_state_updates();

    match wait_for_active_timeline(&mut timeline_state_updates).await {
        ControlFlow::Continue(()) => {}
@@ -57,8 +117,8 @@ pub(super) async fn connection_manager_loop_step(
    }

    let id = TenantTimelineId {
-        tenant_id: connection_manager_state.timeline.tenant_id,
-        timeline_id: connection_manager_state.timeline.timeline_id,
+        tenant_id: walreceiver_state.timeline.tenant_id,
+        timeline_id: walreceiver_state.timeline.timeline_id,
    };

    // Subscribe to the broker updates. Stream shares underlying TCP connection
@@ -68,7 +128,7 @@ pub(super) async fn connection_manager_loop_step(
    info!("Subscribed for broker timeline updates");

    loop {
-        let time_until_next_retry = connection_manager_state.time_until_next_retry();
+        let time_until_next_retry = walreceiver_state.time_until_next_retry();

        // These things are happening concurrently:
        //
@@ -81,12 +141,12 @@ pub(super) async fn connection_manager_loop_step(
        //  - timeline state changes to something that does not allow walreceiver to run concurrently
        select! {
            Some(wal_connection_update) = async {
-                match connection_manager_state.wal_connection.as_mut() {
+                match walreceiver_state.wal_connection.as_mut() {
                    Some(wal_connection) => Some(wal_connection.connection_task.next_task_event().await),
                    None => None,
                }
            } => {
-                let wal_connection = connection_manager_state.wal_connection.as_mut()
+                let wal_connection = walreceiver_state.wal_connection.as_mut()
                    .expect("Should have a connection, as checked by the corresponding select! guard");
                match wal_connection_update {
                    TaskEvent::Update(TaskStateUpdate::Started) => {},
@@ -96,7 +156,7 @@ pub(super) async fn connection_manager_loop_step(
                            // from this safekeeper. This is good enough to clean unsuccessful
                            // retries history and allow reconnecting to this safekeeper without
                            // sleeping for a long time.
-                            connection_manager_state.wal_connection_retries.remove(&wal_connection.sk_id);
+                            walreceiver_state.wal_connection_retries.remove(&wal_connection.sk_id);
                        }
                        wal_connection.status = new_status;
                    }
@@ -105,7 +165,7 @@ pub(super) async fn connection_manager_loop_step(
                            Ok(()) => debug!("WAL receiving task finished"),
                            Err(e) => error!("wal receiver task finished with an error: {e:?}"),
                        }
-                        connection_manager_state.drop_old_connection(false).await;
+                        walreceiver_state.drop_old_connection(false).await;
                    },
                }
            },
@@ -113,7 +173,7 @@ pub(super) async fn connection_manager_loop_step(
            // Got a new update from the broker
            broker_update = broker_subscription.message() => {
                match broker_update {
-                    Ok(Some(broker_update)) => connection_manager_state.register_timeline_update(broker_update),
+                    Ok(Some(broker_update)) => walreceiver_state.register_timeline_update(broker_update),
                    Err(e) => {
                        error!("broker subscription failed: {e}");
                        return ControlFlow::Continue(());
@@ -127,12 +187,12 @@ pub(super) async fn connection_manager_loop_step(

            new_event = async {
                loop {
-                    if connection_manager_state.timeline.current_state() == TimelineState::Loading {
+                    if walreceiver_state.timeline.current_state() == TimelineState::Loading {
                        warn!("wal connection manager should only be launched after timeline has become active");
                    }
                    match timeline_state_updates.changed().await {
                        Ok(()) => {
-                            let new_state = connection_manager_state.timeline.current_state();
+                            let new_state = walreceiver_state.timeline.current_state();
                            match new_state {
                                // we're already active as walreceiver, no need to reactivate
                                TimelineState::Active => continue,
@@ -174,10 +234,14 @@ pub(super) async fn connection_manager_loop_step(
            } => debug!("Waking up for the next retry after waiting for {time_until_next_retry:?}"),
        }

-        if let Some(new_candidate) = connection_manager_state.next_connection_candidate() {
+        if let Some(new_candidate) = walreceiver_state.next_connection_candidate() {
            info!("Switching to new connection candidate: {new_candidate:?}");
-            connection_manager_state
-                .change_connection(new_candidate, ctx)
+            walreceiver_state
+                .change_connection(
+                    new_candidate.safekeeper_id,
+                    new_candidate.wal_source_connconf,
+                    ctx,
+                )
                .await
        }
    }
@@ -254,17 +318,25 @@ const WALCONNECTION_RETRY_MAX_BACKOFF_SECONDS: f64 = 15.0;
 const WALCONNECTION_RETRY_BACKOFF_MULTIPLIER: f64 = 1.5;

 /// All data that's needed to run endless broker loop and keep the WAL streaming connection alive, if possible.
-pub(super) struct ConnectionManagerState {
+struct WalreceiverState {
    id: TenantTimelineId,
+
    /// Use pageserver data about the timeline to filter out some of the safekeepers.
    timeline: Arc<Timeline>,
-    conf: WalReceiverConf,
+    /// The timeout on the connection to safekeeper for WAL streaming.
+    wal_connect_timeout: Duration,
+    /// The timeout to use to determine when the current connection is "stale" and reconnect to the other one.
+    lagging_wal_timeout: Duration,
+    /// The Lsn lag to use to determine when the current connection is lagging to much behind and reconnect to the other one.
+    max_lsn_wal_lag: NonZeroU64,
    /// Current connection to safekeeper for WAL streaming.
    wal_connection: Option<WalConnection>,
    /// Info about retries and unsuccessful attempts to connect to safekeepers.
    wal_connection_retries: HashMap<NodeId, RetryInfo>,
    /// Data about all timelines, available for connection, fetched from storage broker, grouped by their corresponding safekeeper node id.
    wal_stream_candidates: HashMap<NodeId, BrokerSkTimeline>,
+    auth_token: Option<Arc<String>>,
+    availability_zone: Option<String>,
 }

 /// Current connection data.
@@ -274,8 +346,6 @@ struct WalConnection {
    started_at: NaiveDateTime,
    /// Current safekeeper pageserver is connected to for WAL streaming.
    sk_id: NodeId,
-    /// Availability zone of the safekeeper.
-    availability_zone: Option<String>,
    /// Status of the connection.
    status: WalConnectionStatus,
    /// WAL streaming task handle.
@@ -307,8 +377,15 @@ struct BrokerSkTimeline {
    latest_update: NaiveDateTime,
 }

-impl ConnectionManagerState {
-    pub(super) fn new(timeline: Arc<Timeline>, conf: WalReceiverConf) -> Self {
+impl WalreceiverState {
+    fn new(
+        timeline: Arc<Timeline>,
+        wal_connect_timeout: Duration,
+        lagging_wal_timeout: Duration,
+        max_lsn_wal_lag: NonZeroU64,
+        auth_token: Option<Arc<String>>,
+        availability_zone: Option<String>,
+    ) -> Self {
        let id = TenantTimelineId {
            tenant_id: timeline.tenant_id,
            timeline_id: timeline.timeline_id,
@@ -316,19 +393,28 @@ impl ConnectionManagerState {
        Self {
            id,
            timeline,
-            conf,
+            wal_connect_timeout,
+            lagging_wal_timeout,
+            max_lsn_wal_lag,
            wal_connection: None,
            wal_stream_candidates: HashMap::new(),
            wal_connection_retries: HashMap::new(),
+            auth_token,
+            availability_zone,
        }
    }

    /// Shuts down the current connection (if any) and immediately starts another one with the given connection string.
-    async fn change_connection(&mut self, new_sk: NewWalConnectionCandidate, ctx: &RequestContext) {
+    async fn change_connection(
+        &mut self,
+        new_sk_id: NodeId,
+        new_wal_source_connconf: PgConnectionConfig,
+        ctx: &RequestContext,
+    ) {
        self.drop_old_connection(true).await;

        let id = self.id;
-        let connect_timeout = self.conf.wal_connect_timeout;
+        let connect_timeout = self.wal_connect_timeout;
        let timeline = Arc::clone(&self.timeline);
        let ctx = ctx.detached_child(
            TaskKind::WalReceiverConnectionHandler,
@@ -338,7 +424,7 @@ impl ConnectionManagerState {
            async move {
                super::walreceiver_connection::handle_walreceiver_connection(
                    timeline,
-                    new_sk.wal_source_connconf,
+                    new_wal_source_connconf,
                    events_sender,
                    cancellation,
                    connect_timeout,
@@ -347,16 +433,13 @@ impl ConnectionManagerState {
                .await
                .context("walreceiver connection handling failure")
            }
-            .instrument(
-                info_span!("walreceiver_connection", id = %id, node_id = %new_sk.safekeeper_id),
-            )
+            .instrument(info_span!("walreceiver_connection", id = %id, node_id = %new_sk_id))
        });

        let now = Utc::now().naive_utc();
        self.wal_connection = Some(WalConnection {
            started_at: now,
-            sk_id: new_sk.safekeeper_id,
-            availability_zone: new_sk.availability_zone,
+            sk_id: new_sk_id,
            status: WalConnectionStatus {
                is_connected: false,
                has_processed_wal: false,
@@ -463,7 +546,6 @@ impl ConnectionManagerState {
    /// * if connected safekeeper is not present, pick the candidate
    /// * if we haven't received any updates for some time, pick the candidate
    /// * if the candidate commit_lsn is much higher than the current one, pick the candidate
-    /// * if the candidate commit_lsn is same, but candidate is located in the same AZ as the pageserver, pick the candidate
    /// * if connected safekeeper stopped sending us new WAL which is available on other safekeeper, pick the candidate
    ///
    /// This way we ensure to keep up with the most up-to-date safekeeper and don't try to jump from one safekeeper to another too frequently.
@@ -477,24 +559,22 @@ impl ConnectionManagerState {

                let (new_sk_id, new_safekeeper_broker_data, new_wal_source_connconf) =
                    self.select_connection_candidate(Some(connected_sk_node))?;
-                let new_availability_zone = new_safekeeper_broker_data.availability_zone.clone();

                let now = Utc::now().naive_utc();
                if let Ok(latest_interaciton) =
                    (now - existing_wal_connection.status.latest_connection_update).to_std()
                {
                    // Drop connection if we haven't received keepalive message for a while.
-                    if latest_interaciton > self.conf.wal_connect_timeout {
+                    if latest_interaciton > self.wal_connect_timeout {
                        return Some(NewWalConnectionCandidate {
                            safekeeper_id: new_sk_id,
                            wal_source_connconf: new_wal_source_connconf,
-                            availability_zone: new_availability_zone,
                            reason: ReconnectReason::NoKeepAlives {
                                last_keep_alive: Some(
                                    existing_wal_connection.status.latest_connection_update,
                                ),
                                check_time: now,
-                                threshold: self.conf.wal_connect_timeout,
+                                threshold: self.wal_connect_timeout,
                            },
                        });
                    }
@@ -510,32 +590,17 @@ impl ConnectionManagerState {
                    // Check if the new candidate has much more WAL than the current one.
                    match new_commit_lsn.0.checked_sub(current_commit_lsn.0) {
                        Some(new_sk_lsn_advantage) => {
-                            if new_sk_lsn_advantage >= self.conf.max_lsn_wal_lag.get() {
+                            if new_sk_lsn_advantage >= self.max_lsn_wal_lag.get() {
                                return Some(NewWalConnectionCandidate {
                                    safekeeper_id: new_sk_id,
                                    wal_source_connconf: new_wal_source_connconf,
-                                    availability_zone: new_availability_zone,
                                    reason: ReconnectReason::LaggingWal {
                                        current_commit_lsn,
                                        new_commit_lsn,
-                                        threshold: self.conf.max_lsn_wal_lag,
+                                        threshold: self.max_lsn_wal_lag,
                                    },
                                });
                            }
-                            // If we have a candidate with the same commit_lsn as the current one, which is in the same AZ as pageserver,
-                            // and the current one is not, switch to the new one.
-                            if self.conf.availability_zone.is_some()
-                                && existing_wal_connection.availability_zone
-                                    != self.conf.availability_zone
-                                && self.conf.availability_zone == new_availability_zone
-                            {
-                                return Some(NewWalConnectionCandidate {
-                                    safekeeper_id: new_sk_id,
-                                    availability_zone: new_availability_zone,
-                                    wal_source_connconf: new_wal_source_connconf,
-                                    reason: ReconnectReason::SwitchAvailabilityZone,
-                                });
-                            }
                        }
                        None => debug!(
                            "Best SK candidate has its commit_lsn behind connected SK's commit_lsn"
@@ -598,12 +663,11 @@ impl ConnectionManagerState {
                if let Some(waiting_for_new_lsn_since) = waiting_for_new_lsn_since {
                    if let Ok(waiting_for_new_wal) = (now - waiting_for_new_lsn_since).to_std() {
                        if candidate_commit_lsn > current_commit_lsn
-                            && waiting_for_new_wal > self.conf.lagging_wal_timeout
+                            && waiting_for_new_wal > self.lagging_wal_timeout
                        {
                            return Some(NewWalConnectionCandidate {
                                safekeeper_id: new_sk_id,
                                wal_source_connconf: new_wal_source_connconf,
-                                availability_zone: new_availability_zone,
                                reason: ReconnectReason::NoWalTimeout {
                                    current_lsn,
                                    current_commit_lsn,
@@ -612,7 +676,7 @@ impl ConnectionManagerState {
                                        existing_wal_connection.status.latest_wal_update,
                                    ),
                                    check_time: now,
-                                    threshold: self.conf.lagging_wal_timeout,
+                                    threshold: self.lagging_wal_timeout,
                                },
                            });
                        }
@@ -622,11 +686,10 @@ impl ConnectionManagerState {
                self.wal_connection.as_mut().unwrap().discovered_new_wal = discovered_new_wal;
            }
            None => {
-                let (new_sk_id, new_safekeeper_broker_data, new_wal_source_connconf) =
+                let (new_sk_id, _, new_wal_source_connconf) =
                    self.select_connection_candidate(None)?;
                return Some(NewWalConnectionCandidate {
                    safekeeper_id: new_sk_id,
-                    availability_zone: new_safekeeper_broker_data.availability_zone.clone(),
                    wal_source_connconf: new_wal_source_connconf,
                    reason: ReconnectReason::NoExistingConnection,
                });
@@ -678,11 +741,11 @@ impl ConnectionManagerState {
                match wal_stream_connection_config(
                    self.id,
                    info.safekeeper_connstr.as_ref(),
-                    match &self.conf.auth_token {
+                    match &self.auth_token {
                        None => None,
                        Some(x) => Some(x),
                    },
-                    self.conf.availability_zone.as_deref(),
+                    self.availability_zone.as_deref(),
                ) {
                    Ok(connstr) => Some((*sk_id, info, connstr)),
                    Err(e) => {
@@ -696,7 +759,7 @@ impl ConnectionManagerState {
    /// Remove candidates which haven't sent broker updates for a while.
    fn cleanup_old_candidates(&mut self) {
        let mut node_ids_to_remove = Vec::with_capacity(self.wal_stream_candidates.len());
-        let lagging_wal_timeout = self.conf.lagging_wal_timeout;
+        let lagging_wal_timeout = self.lagging_wal_timeout;

        self.wal_stream_candidates.retain(|node_id, broker_info| {
            if let Ok(time_since_latest_broker_update) =
@@ -720,7 +783,7 @@ impl ConnectionManagerState {
        }
    }

-    pub(super) async fn shutdown(mut self) {
+    async fn shutdown(mut self) {
        if let Some(wal_connection) = self.wal_connection.take() {
            wal_connection.connection_task.shutdown().await;
        }
@@ -731,7 +794,6 @@ impl ConnectionManagerState {
 struct NewWalConnectionCandidate {
    safekeeper_id: NodeId,
    wal_source_connconf: PgConnectionConfig,
-    availability_zone: Option<String>,
    // This field is used in `derive(Debug)` only.
    #[allow(dead_code)]
    reason: ReconnectReason,
@@ -746,7 +808,6 @@ enum ReconnectReason {
        new_commit_lsn: Lsn,
        threshold: NonZeroU64,
    },
-    SwitchAvailabilityZone,
    NoWalTimeout {
        current_lsn: Lsn,
        current_commit_lsn: Lsn,
@@ -812,7 +873,6 @@ mod tests {
                peer_horizon_lsn: 0,
                local_start_lsn: 0,
                safekeeper_connstr: safekeeper_connstr.to_owned(),
-                availability_zone: None,
            },
            latest_update,
        }
@@ -824,7 +884,7 @@ mod tests {
        let mut state = dummy_state(&harness).await;
        let now = Utc::now().naive_utc();

-        let lagging_wal_timeout = chrono::Duration::from_std(state.conf.lagging_wal_timeout)?;
+        let lagging_wal_timeout = chrono::Duration::from_std(state.lagging_wal_timeout)?;
        let delay_over_threshold = now - lagging_wal_timeout - lagging_wal_timeout;

        state.wal_connection = None;
@@ -835,7 +895,7 @@ mod tests {
            (
                NodeId(3),
                dummy_broker_sk_timeline(
-                    1 + state.conf.max_lsn_wal_lag.get(),
+                    1 + state.max_lsn_wal_lag.get(),
                    "delay_over_threshold",
                    delay_over_threshold,
                ),
@@ -869,11 +929,10 @@ mod tests {
            streaming_lsn: Some(Lsn(current_lsn)),
        };

-        state.conf.max_lsn_wal_lag = NonZeroU64::new(100).unwrap();
+        state.max_lsn_wal_lag = NonZeroU64::new(100).unwrap();
        state.wal_connection = Some(WalConnection {
            started_at: now,
            sk_id: connected_sk_id,
-            availability_zone: None,
            status: connection_status,
            connection_task: TaskHandle::spawn(move |sender, _| async move {
                sender
@@ -887,7 +946,7 @@ mod tests {
            (
                connected_sk_id,
                dummy_broker_sk_timeline(
-                    current_lsn + state.conf.max_lsn_wal_lag.get() * 2,
+                    current_lsn + state.max_lsn_wal_lag.get() * 2,
                    DUMMY_SAFEKEEPER_HOST,
                    now,
                ),
@@ -899,7 +958,7 @@ mod tests {
            (
                NodeId(2),
                dummy_broker_sk_timeline(
-                    current_lsn + state.conf.max_lsn_wal_lag.get() / 2,
+                    current_lsn + state.max_lsn_wal_lag.get() / 2,
                    "not_enough_advanced_lsn",
                    now,
                ),
@@ -924,11 +983,7 @@ mod tests {
        state.wal_connection = None;
        state.wal_stream_candidates = HashMap::from([(
            NodeId(0),
-            dummy_broker_sk_timeline(
-                1 + state.conf.max_lsn_wal_lag.get(),
-                DUMMY_SAFEKEEPER_HOST,
-                now,
-            ),
+            dummy_broker_sk_timeline(1 + state.max_lsn_wal_lag.get(), DUMMY_SAFEKEEPER_HOST, now),
        )]);

        let only_candidate = state
@@ -1026,7 +1081,7 @@ mod tests {
        let now = Utc::now().naive_utc();

        let connected_sk_id = NodeId(0);
-        let new_lsn = Lsn(current_lsn.0 + state.conf.max_lsn_wal_lag.get() + 1);
+        let new_lsn = Lsn(current_lsn.0 + state.max_lsn_wal_lag.get() + 1);

        let connection_status = WalConnectionStatus {
            is_connected: true,
@@ -1040,7 +1095,6 @@ mod tests {
        state.wal_connection = Some(WalConnection {
            started_at: now,
            sk_id: connected_sk_id,
-            availability_zone: None,
            status: connection_status,
            connection_task: TaskHandle::spawn(move |sender, _| async move {
                sender
@@ -1071,7 +1125,7 @@ mod tests {
            ReconnectReason::LaggingWal {
                current_commit_lsn: current_lsn,
                new_commit_lsn: new_lsn,
-                threshold: state.conf.max_lsn_wal_lag
+                threshold: state.max_lsn_wal_lag
            },
            "Should select bigger WAL safekeeper if it starts to lag enough"
        );
@@ -1090,7 +1144,7 @@ mod tests {
        let current_lsn = Lsn(100_000).align();
        let now = Utc::now().naive_utc();

-        let wal_connect_timeout = chrono::Duration::from_std(state.conf.wal_connect_timeout)?;
+        let wal_connect_timeout = chrono::Duration::from_std(state.wal_connect_timeout)?;
        let time_over_threshold =
            Utc::now().naive_utc() - wal_connect_timeout - wal_connect_timeout;

@@ -1106,7 +1160,6 @@ mod tests {
        state.wal_connection = Some(WalConnection {
            started_at: now,
            sk_id: NodeId(1),
-            availability_zone: None,
            status: connection_status,
            connection_task: TaskHandle::spawn(move |sender, _| async move {
                sender
@@ -1133,7 +1186,7 @@ mod tests {
                ..
            } => {
                assert_eq!(last_keep_alive, Some(time_over_threshold));
-                assert_eq!(threshold, state.conf.lagging_wal_timeout);
+                assert_eq!(threshold, state.lagging_wal_timeout);
            }
            unexpected => panic!("Unexpected reason: {unexpected:?}"),
        }
@@ -1153,7 +1206,7 @@ mod tests {
        let new_lsn = Lsn(100_100).align();
        let now = Utc::now().naive_utc();

-        let lagging_wal_timeout = chrono::Duration::from_std(state.conf.lagging_wal_timeout)?;
+        let lagging_wal_timeout = chrono::Duration::from_std(state.lagging_wal_timeout)?;
        let time_over_threshold =
            Utc::now().naive_utc() - lagging_wal_timeout - lagging_wal_timeout;

@@ -1169,7 +1222,6 @@ mod tests {
        state.wal_connection = Some(WalConnection {
            started_at: now,
            sk_id: NodeId(1),
-            availability_zone: None,
            status: connection_status,
            connection_task: TaskHandle::spawn(move |_, _| async move { Ok(()) }),
            discovered_new_wal: Some(NewCommittedWAL {
@@ -1200,7 +1252,7 @@ mod tests {
                assert_eq!(current_commit_lsn, current_lsn);
                assert_eq!(candidate_commit_lsn, new_lsn);
                assert_eq!(last_wal_interaction, Some(time_over_threshold));
-                assert_eq!(threshold, state.conf.lagging_wal_timeout);
+                assert_eq!(threshold, state.lagging_wal_timeout);
            }
            unexpected => panic!("Unexpected reason: {unexpected:?}"),
        }
@@ -1214,99 +1266,27 @@ mod tests {

    const DUMMY_SAFEKEEPER_HOST: &str = "safekeeper_connstr";

-    async fn dummy_state(harness: &TenantHarness<'_>) -> ConnectionManagerState {
+    async fn dummy_state(harness: &TenantHarness<'_>) -> WalreceiverState {
        let (tenant, ctx) = harness.load().await;
        let timeline = tenant
            .create_empty_timeline(TIMELINE_ID, Lsn(0), crate::DEFAULT_PG_VERSION, &ctx)
            .expect("Failed to create an empty timeline for dummy wal connection manager");
        let timeline = timeline.initialize(&ctx).unwrap();

-        ConnectionManagerState {
+        WalreceiverState {
            id: TenantTimelineId {
                tenant_id: harness.tenant_id,
                timeline_id: TIMELINE_ID,
            },
            timeline,
-            conf: WalReceiverConf {
-                wal_connect_timeout: Duration::from_secs(1),
-                lagging_wal_timeout: Duration::from_secs(1),
-                max_lsn_wal_lag: NonZeroU64::new(1024 * 1024).unwrap(),
-                auth_token: None,
-                availability_zone: None,
-            },
+            wal_connect_timeout: Duration::from_secs(1),
+            lagging_wal_timeout: Duration::from_secs(1),
+            max_lsn_wal_lag: NonZeroU64::new(1024 * 1024).unwrap(),
            wal_connection: None,
            wal_stream_candidates: HashMap::new(),
            wal_connection_retries: HashMap::new(),
+            auth_token: None,
+            availability_zone: None,
        }
    }
-
-    #[tokio::test]
-    async fn switch_to_same_availability_zone() -> anyhow::Result<()> {
-        // Pageserver and one of safekeepers will be in the same availability zone
-        // and pageserver should prefer to connect to it.
-        let test_az = Some("test_az".to_owned());
-
-        let harness = TenantHarness::create("switch_to_same_availability_zone")?;
-        let mut state = dummy_state(&harness).await;
-        state.conf.availability_zone = test_az.clone();
-        let current_lsn = Lsn(100_000).align();
-        let now = Utc::now().naive_utc();
-
-        let connected_sk_id = NodeId(0);
-
-        let connection_status = WalConnectionStatus {
-            is_connected: true,
-            has_processed_wal: true,
-            latest_connection_update: now,
-            latest_wal_update: now,
-            commit_lsn: Some(current_lsn),
-            streaming_lsn: Some(current_lsn),
-        };
-
-        state.wal_connection = Some(WalConnection {
-            started_at: now,
-            sk_id: connected_sk_id,
-            availability_zone: None,
-            status: connection_status,
-            connection_task: TaskHandle::spawn(move |sender, _| async move {
-                sender
-                    .send(TaskStateUpdate::Progress(connection_status))
-                    .ok();
-                Ok(())
-            }),
-            discovered_new_wal: None,
-        });
-
-        // We have another safekeeper with the same commit_lsn, and it have the same availability zone as
-        // the current pageserver.
-        let mut same_az_sk = dummy_broker_sk_timeline(current_lsn.0, "same_az", now);
-        same_az_sk.timeline.availability_zone = test_az.clone();
-
-        state.wal_stream_candidates = HashMap::from([
-            (
-                connected_sk_id,
-                dummy_broker_sk_timeline(current_lsn.0, DUMMY_SAFEKEEPER_HOST, now),
-            ),
-            (NodeId(1), same_az_sk),
-        ]);
-
-        // We expect that pageserver will switch to the safekeeper in the same availability zone,
-        // even if it has the same commit_lsn.
-        let next_candidate = state.next_connection_candidate().expect(
-            "Expected one candidate selected out of multiple valid data options, but got none",
-        );
-
-        assert_eq!(next_candidate.safekeeper_id, NodeId(1));
-        assert_eq!(
-            next_candidate.reason,
-            ReconnectReason::SwitchAvailabilityZone,
-            "Should switch to the safekeeper in the same availability zone, if it has the same commit_lsn"
-        );
-        assert_eq!(
-            next_candidate.wal_source_connconf.host(),
-            &Host::Domain("same_az".to_owned())
-        );
-
-        Ok(())
-    }
 }
--- a/pageserver/src/tenant/timeline/walreceiver/walreceiver_connection.rs
+++ b/pageserver/src/tenant/timeline/walreceiver/walreceiver_connection.rs
@@ -37,12 +37,12 @@ use crate::{
 use postgres_backend::is_expected_io_error;
 use postgres_connection::PgConnectionConfig;
 use postgres_ffi::waldecoder::WalStreamDecoder;
-use pq_proto::PageserverFeedback;
+use pq_proto::ReplicationFeedback;
 use utils::lsn::Lsn;

 /// Status of the connection.
 #[derive(Debug, Clone, Copy)]
-pub(super) struct WalConnectionStatus {
+pub struct WalConnectionStatus {
    /// If we were able to initiate a postgres connection, this means that safekeeper process is at least running.
    pub is_connected: bool,
    /// Defines a healthy connection as one on which pageserver received WAL from safekeeper
@@ -60,7 +60,7 @@ pub(super) struct WalConnectionStatus {

 /// Open a connection to the given safekeeper and receive WAL, sending back progress
 /// messages as we go.
-pub(super) async fn handle_walreceiver_connection(
+pub async fn handle_walreceiver_connection(
    timeline: Arc<Timeline>,
    wal_source_connconf: PgConnectionConfig,
    events_sender: watch::Sender<TaskStateUpdate<WalConnectionStatus>>,
@@ -319,12 +319,12 @@ pub(super) async fn handle_walreceiver_connection(
                timeline.get_remote_consistent_lsn().unwrap_or(Lsn(0));

            // The last LSN we processed. It is not guaranteed to survive pageserver crash.
-            let last_received_lsn = u64::from(last_lsn);
+            let write_lsn = u64::from(last_lsn);
            // `disk_consistent_lsn` is the LSN at which page server guarantees local persistence of all received data
-            let disk_consistent_lsn = u64::from(timeline.get_disk_consistent_lsn());
+            let flush_lsn = u64::from(timeline.get_disk_consistent_lsn());
            // The last LSN that is synced to remote storage and is guaranteed to survive pageserver crash
            // Used by safekeepers to remove WAL preceding `remote_consistent_lsn`.
-            let remote_consistent_lsn = u64::from(timeline_remote_consistent_lsn);
+            let apply_lsn = u64::from(timeline_remote_consistent_lsn);
            let ts = SystemTime::now();

            // Update the status about what we just received. This is shown in the mgmt API.
@@ -343,12 +343,12 @@ pub(super) async fn handle_walreceiver_connection(
            let (timeline_logical_size, _) = timeline
                .get_current_logical_size(&ctx)
                .context("Status update creation failed to get current logical size")?;
-            let status_update = PageserverFeedback {
+            let status_update = ReplicationFeedback {
                current_timeline_size: timeline_logical_size,
-                last_received_lsn,
-                disk_consistent_lsn,
-                remote_consistent_lsn,
-                replytime: ts,
+                ps_writelsn: write_lsn,
+                ps_flushlsn: flush_lsn,
+                ps_applylsn: apply_lsn,
+                ps_replytime: ts,
            };

            debug!("neon_status_update {status_update:?}");
--- a/pgxn/neon/walproposer.c
+++ b/pgxn/neon/walproposer.c
@@ -1872,9 +1872,9 @@ RecvAppendResponses(Safekeeper *sk)
 	return sk->state == SS_ACTIVE;
 }

-/* Parse a PageserverFeedback message, or the PageserverFeedback part of an AppendResponse */
+/* Parse a ReplicationFeedback message, or the ReplicationFeedback part of an AppendResponse */
 void
-ParsePageserverFeedbackMessage(StringInfo reply_message, PageserverFeedback * rf)
+ParseReplicationFeedbackMessage(StringInfo reply_message, ReplicationFeedback * rf)
 {
 	uint8		nkeys;
 	int			i;
@@ -1892,45 +1892,45 @@ ParsePageserverFeedbackMessage(StringInfo reply_message, PageserverFeedback * rf
 			pq_getmsgint(reply_message, sizeof(int32));
 			/* read value length */
 			rf->currentClusterSize = pq_getmsgint64(reply_message);
-			elog(DEBUG2, "ParsePageserverFeedbackMessage: current_timeline_size %lu",
+			elog(DEBUG2, "ParseReplicationFeedbackMessage: current_timeline_size %lu",
 				 rf->currentClusterSize);
 		}
-		else if ((strcmp(key, "ps_writelsn") == 0) || (strcmp(key, "last_received_lsn") == 0))
+		else if (strcmp(key, "ps_writelsn") == 0)
 		{
 			pq_getmsgint(reply_message, sizeof(int32));
 			/* read value length */
-			rf->last_received_lsn = pq_getmsgint64(reply_message);
-			elog(DEBUG2, "ParsePageserverFeedbackMessage: last_received_lsn %X/%X",
-				 LSN_FORMAT_ARGS(rf->last_received_lsn));
+			rf->ps_writelsn = pq_getmsgint64(reply_message);
+			elog(DEBUG2, "ParseReplicationFeedbackMessage: ps_writelsn %X/%X",
+				 LSN_FORMAT_ARGS(rf->ps_writelsn));
 		}
-		else if ((strcmp(key, "ps_flushlsn") == 0) || (strcmp(key, "disk_consistent_lsn") == 0))
+		else if (strcmp(key, "ps_flushlsn") == 0)
 		{
 			pq_getmsgint(reply_message, sizeof(int32));
 			/* read value length */
-			rf->disk_consistent_lsn = pq_getmsgint64(reply_message);
-			elog(DEBUG2, "ParsePageserverFeedbackMessage: disk_consistent_lsn %X/%X",
-				 LSN_FORMAT_ARGS(rf->disk_consistent_lsn));
+			rf->ps_flushlsn = pq_getmsgint64(reply_message);
+			elog(DEBUG2, "ParseReplicationFeedbackMessage: ps_flushlsn %X/%X",
+				 LSN_FORMAT_ARGS(rf->ps_flushlsn));
 		}
-		else if ((strcmp(key, "ps_applylsn") == 0) || (strcmp(key, "remote_consistent_lsn") == 0))
+		else if (strcmp(key, "ps_applylsn") == 0)
 		{
 			pq_getmsgint(reply_message, sizeof(int32));
 			/* read value length */
-			rf->remote_consistent_lsn = pq_getmsgint64(reply_message);
-			elog(DEBUG2, "ParsePageserverFeedbackMessage: remote_consistent_lsn %X/%X",
-				 LSN_FORMAT_ARGS(rf->remote_consistent_lsn));
+			rf->ps_applylsn = pq_getmsgint64(reply_message);
+			elog(DEBUG2, "ParseReplicationFeedbackMessage: ps_applylsn %X/%X",
+				 LSN_FORMAT_ARGS(rf->ps_applylsn));
 		}
-		else if ((strcmp(key, "ps_replytime") == 0) || (strcmp(key, "replytime") == 0))
+		else if (strcmp(key, "ps_replytime") == 0)
 		{
 			pq_getmsgint(reply_message, sizeof(int32));
 			/* read value length */
-			rf->replytime = pq_getmsgint64(reply_message);
+			rf->ps_replytime = pq_getmsgint64(reply_message);
 			{
 				char	   *replyTimeStr;

 				/* Copy because timestamptz_to_str returns a static buffer */
-				replyTimeStr = pstrdup(timestamptz_to_str(rf->replytime));
-				elog(DEBUG2, "ParsePageserverFeedbackMessage: replytime %lu reply_time: %s",
-					 rf->replytime, replyTimeStr);
+				replyTimeStr = pstrdup(timestamptz_to_str(rf->ps_replytime));
+				elog(DEBUG2, "ParseReplicationFeedbackMessage: ps_replytime %lu reply_time: %s",
+					 rf->ps_replytime, replyTimeStr);

 				pfree(replyTimeStr);
 			}
@@ -1944,7 +1944,7 @@ ParsePageserverFeedbackMessage(StringInfo reply_message, PageserverFeedback * rf
 			 * Skip unknown keys to support backward compatibile protocol
 			 * changes
 			 */
-			elog(LOG, "ParsePageserverFeedbackMessage: unknown key: %s len %d", key, len);
+			elog(LOG, "ParseReplicationFeedbackMessage: unknown key: %s len %d", key, len);
 			pq_getmsgbytes(reply_message, len);
 		};
 	}
@@ -2024,7 +2024,7 @@ GetAcknowledgedByQuorumWALPosition(void)
 }

 /*
- * WalproposerShmemSize --- report amount of shared memory space needed
+ * ReplicationFeedbackShmemSize --- report amount of shared memory space needed
 */
 Size
 WalproposerShmemSize(void)
@@ -2054,10 +2054,10 @@ WalproposerShmemInit(void)
 }

 void
-replication_feedback_set(PageserverFeedback * rf)
+replication_feedback_set(ReplicationFeedback * rf)
 {
 	SpinLockAcquire(&walprop_shared->mutex);
-	memcpy(&walprop_shared->feedback, rf, sizeof(PageserverFeedback));
+	memcpy(&walprop_shared->feedback, rf, sizeof(ReplicationFeedback));
 	SpinLockRelease(&walprop_shared->mutex);
 }

@@ -2065,43 +2065,43 @@ void
 replication_feedback_get_lsns(XLogRecPtr *writeLsn, XLogRecPtr *flushLsn, XLogRecPtr *applyLsn)
 {
 	SpinLockAcquire(&walprop_shared->mutex);
-	*writeLsn = walprop_shared->feedback.last_received_lsn;
-	*flushLsn = walprop_shared->feedback.disk_consistent_lsn;
-	*applyLsn = walprop_shared->feedback.remote_consistent_lsn;
+	*writeLsn = walprop_shared->feedback.ps_writelsn;
+	*flushLsn = walprop_shared->feedback.ps_flushlsn;
+	*applyLsn = walprop_shared->feedback.ps_applylsn;
 	SpinLockRelease(&walprop_shared->mutex);
 }

 /*
- * Get PageserverFeedback fields from the most advanced safekeeper
+ * Get ReplicationFeedback fields from the most advanced safekeeper
 */
 static void
-GetLatestNeonFeedback(PageserverFeedback * rf)
+GetLatestNeonFeedback(ReplicationFeedback * rf)
 {
 	int			latest_safekeeper = 0;
-	XLogRecPtr	last_received_lsn = InvalidXLogRecPtr;
+	XLogRecPtr	ps_writelsn = InvalidXLogRecPtr;

 	for (int i = 0; i < n_safekeepers; i++)
 	{
-		if (safekeeper[i].appendResponse.rf.last_received_lsn > last_received_lsn)
+		if (safekeeper[i].appendResponse.rf.ps_writelsn > ps_writelsn)
 		{
 			latest_safekeeper = i;
-			last_received_lsn = safekeeper[i].appendResponse.rf.last_received_lsn;
+			ps_writelsn = safekeeper[i].appendResponse.rf.ps_writelsn;
 		}
 	}

 	rf->currentClusterSize = safekeeper[latest_safekeeper].appendResponse.rf.currentClusterSize;
-	rf->last_received_lsn = safekeeper[latest_safekeeper].appendResponse.rf.last_received_lsn;
-	rf->disk_consistent_lsn = safekeeper[latest_safekeeper].appendResponse.rf.disk_consistent_lsn;
-	rf->remote_consistent_lsn = safekeeper[latest_safekeeper].appendResponse.rf.remote_consistent_lsn;
-	rf->replytime = safekeeper[latest_safekeeper].appendResponse.rf.replytime;
+	rf->ps_writelsn = safekeeper[latest_safekeeper].appendResponse.rf.ps_writelsn;
+	rf->ps_flushlsn = safekeeper[latest_safekeeper].appendResponse.rf.ps_flushlsn;
+	rf->ps_applylsn = safekeeper[latest_safekeeper].appendResponse.rf.ps_applylsn;
+	rf->ps_replytime = safekeeper[latest_safekeeper].appendResponse.rf.ps_replytime;

 	elog(DEBUG2, "GetLatestNeonFeedback: currentClusterSize %lu,"
-		 " last_received_lsn %X/%X, disk_consistent_lsn %X/%X, remote_consistent_lsn %X/%X, replytime %lu",
+		 " ps_writelsn %X/%X, ps_flushlsn %X/%X, ps_applylsn %X/%X, ps_replytime %lu",
 		 rf->currentClusterSize,
-		 LSN_FORMAT_ARGS(rf->last_received_lsn),
-		 LSN_FORMAT_ARGS(rf->disk_consistent_lsn),
-		 LSN_FORMAT_ARGS(rf->remote_consistent_lsn),
-		 rf->replytime);
+		 LSN_FORMAT_ARGS(rf->ps_writelsn),
+		 LSN_FORMAT_ARGS(rf->ps_flushlsn),
+		 LSN_FORMAT_ARGS(rf->ps_applylsn),
+		 rf->ps_replytime);

 	replication_feedback_set(rf);
 }
@@ -2115,16 +2115,16 @@ HandleSafekeeperResponse(void)
 	XLogRecPtr	minFlushLsn;

 	minQuorumLsn = GetAcknowledgedByQuorumWALPosition();
-	diskConsistentLsn = quorumFeedback.rf.disk_consistent_lsn;
+	diskConsistentLsn = quorumFeedback.rf.ps_flushlsn;

 	if (!syncSafekeepers)
 	{
-		/* Get PageserverFeedback fields from the most advanced safekeeper */
+		/* Get ReplicationFeedback fields from the most advanced safekeeper */
 		GetLatestNeonFeedback(&quorumFeedback.rf);
 		SetZenithCurrentClusterSize(quorumFeedback.rf.currentClusterSize);
 	}

-	if (minQuorumLsn > quorumFeedback.flushLsn || diskConsistentLsn != quorumFeedback.rf.disk_consistent_lsn)
+	if (minQuorumLsn > quorumFeedback.flushLsn || diskConsistentLsn != quorumFeedback.rf.ps_flushlsn)
 	{

 		if (minQuorumLsn > quorumFeedback.flushLsn)
@@ -2142,7 +2142,7 @@ HandleSafekeeperResponse(void)
 			 * apply_lsn - This is what processed and durably saved at*
 			 * pageserver.
 			 */
-								quorumFeedback.rf.disk_consistent_lsn,
+								quorumFeedback.rf.ps_flushlsn,
 								GetCurrentTimestamp(), false);
 	}

@@ -2326,7 +2326,7 @@ AsyncReadMessage(Safekeeper *sk, AcceptorProposerMessage * anymsg)
 				msg->hs.xmin.value = pq_getmsgint64_le(&s);
 				msg->hs.catalog_xmin.value = pq_getmsgint64_le(&s);
 				if (buf_size > APPENDRESPONSE_FIXEDPART_SIZE)
-					ParsePageserverFeedbackMessage(&s, &msg->rf);
+					ParseReplicationFeedbackMessage(&s, &msg->rf);
 				pq_getmsgend(&s);
 				return true;
 			}
@@ -2462,7 +2462,7 @@ backpressure_lag_impl(void)
 		replication_feedback_get_lsns(&writePtr, &flushPtr, &applyPtr);
 #define MB ((XLogRecPtr)1024 * 1024)

-		elog(DEBUG2, "current flushLsn %X/%X PageserverFeedback: write %X/%X flush %X/%X apply %X/%X",
+		elog(DEBUG2, "current flushLsn %X/%X ReplicationFeedback: write %X/%X flush %X/%X apply %X/%X",
 			 LSN_FORMAT_ARGS(myFlushLsn),
 			 LSN_FORMAT_ARGS(writePtr),
 			 LSN_FORMAT_ARGS(flushPtr),
--- a/pgxn/neon/walproposer.h
+++ b/pgxn/neon/walproposer.h
@@ -280,21 +280,21 @@ typedef struct HotStandbyFeedback
 	FullTransactionId catalog_xmin;
 }			HotStandbyFeedback;

-typedef struct PageserverFeedback
+typedef struct ReplicationFeedback
 {
 	/* current size of the timeline on pageserver */
 	uint64		currentClusterSize;
 	/* standby_status_update fields that safekeeper received from pageserver */
-	XLogRecPtr	last_received_lsn;
-	XLogRecPtr	disk_consistent_lsn;
-	XLogRecPtr	remote_consistent_lsn;
-	TimestampTz replytime;
-}			PageserverFeedback;
+	XLogRecPtr	ps_writelsn;
+	XLogRecPtr	ps_flushlsn;
+	XLogRecPtr	ps_applylsn;
+	TimestampTz ps_replytime;
+}			ReplicationFeedback;

 typedef struct WalproposerShmemState
 {
 	slock_t		mutex;
-	PageserverFeedback feedback;
+	ReplicationFeedback feedback;
 	term_t		mineLastElectedTerm;
 	pg_atomic_uint64 backpressureThrottlingTime;
 }			WalproposerShmemState;
@@ -320,10 +320,10 @@ typedef struct AppendResponse
 	/* Feedback recieved from pageserver includes standby_status_update fields */
 	/* and custom neon feedback. */
 	/* This part of the message is extensible. */
-	PageserverFeedback rf;
+	ReplicationFeedback rf;
 }			AppendResponse;

-/*  PageserverFeedback is extensible part of the message that is parsed separately */
+/*  ReplicationFeedback is extensible part of the message that is parsed separately */
 /*  Other fields are fixed part */
 #define APPENDRESPONSE_FIXEDPART_SIZE offsetof(AppendResponse, rf)

@@ -383,13 +383,13 @@ extern void WalProposerSync(int argc, char *argv[]);
 extern void WalProposerMain(Datum main_arg);
 extern void WalProposerBroadcast(XLogRecPtr startpos, XLogRecPtr endpos);
 extern void WalProposerPoll(void);
-extern void ParsePageserverFeedbackMessage(StringInfo reply_message,
-											PageserverFeedback *rf);
+extern void ParseReplicationFeedbackMessage(StringInfo reply_message,
+											ReplicationFeedback *rf);
 extern void StartProposerReplication(StartReplicationCmd *cmd);

 extern Size WalproposerShmemSize(void);
 extern bool WalproposerShmemInit(void);
-extern void replication_feedback_set(PageserverFeedback *rf);
+extern void replication_feedback_set(ReplicationFeedback *rf);
 extern void replication_feedback_get_lsns(XLogRecPtr *writeLsn, XLogRecPtr *flushLsn, XLogRecPtr *applyLsn);

 /* libpqwalproposer hooks & helper type */
--- a/pgxn/neon_walredo/seccomp.c
+++ b/pgxn/neon_walredo/seccomp.c
@@ -9,14 +9,6 @@
 * To prevent this, it has been decided to limit possible interactions
 * with the outside world using the Secure Computing BPF mode.
 *
- * This code is intended to support both x86_64 and aarch64. The latter
- * doesn't implement some syscalls like open and select. We allow both
- * select (absent on aarch64) and pselect6 (present on both architectures)
- * We call select(2) through libc, and the libc wrapper calls select or pselect6
- * depending on the architecture. You can check which syscalls are present on
- * different architectures with the `scmp_sys_resolver` tool from the
- * seccomp package.
- *
 * We use this mode to disable all syscalls not in the allowlist. This
 * approach has its pros & cons:
 *
@@ -81,6 +73,8 @@
 *    I suspect that certain libc functions might involve slightly
 *    different syscalls, e.g. select/pselect6/pselect6_time64/whatever.
 *
+ *  - Test on any arch other than amd64 to see if it works there.
+ *
 *-------------------------------------------------------------------------
 */

@@ -128,10 +122,9 @@ seccomp_load_rules(PgSeccompRule *rules, int count)

 	/*
 	 * First, check that open of a well-known file works.
-	 * XXX: We use raw syscall() to call the very openat() which is
-	 * present both on x86_64 and on aarch64.
+	 * XXX: We use raw syscall() to call the very open().
 	 */
-	fd = syscall(SCMP_SYS(openat), AT_FDCWD, "/dev/null", O_RDONLY, 0);
+	fd = syscall(SCMP_SYS(open), "/dev/null", O_RDONLY, 0);
 	if (seccomp_test_sighandler_done)
 		ereport(FATAL,
 				(errcode(ERRCODE_SYSTEM_ERROR),
@@ -142,15 +135,15 @@ seccomp_load_rules(PgSeccompRule *rules, int count)
 				 errmsg("seccomp: could not open /dev/null for seccomp testing: %m")));
 	close((int) fd);

-	/* Set a trap on openat() to test seccomp bpf */
-	rule = PG_SCMP(openat, SCMP_ACT_TRAP);
+	/* Set a trap on open() to test seccomp bpf */
+	rule = PG_SCMP(open, SCMP_ACT_TRAP);
 	if (do_seccomp_load_rules(&rule, 1, SCMP_ACT_ALLOW) != 0)
 		ereport(FATAL,
 				(errcode(ERRCODE_SYSTEM_ERROR),
 				 errmsg("seccomp: could not load test trap")));

-	/* Finally, check that openat() now raises SIGSYS */
-	(void) syscall(SCMP_SYS(openat), AT_FDCWD, "/dev/null", O_RDONLY, 0);
+	/* Finally, check that open() now raises SIGSYS */
+	(void) syscall(SCMP_SYS(open), "/dev/null", O_RDONLY, 0);
 	if (!seccomp_test_sighandler_done)
 		ereport(FATAL,
 				(errcode(ERRCODE_SYSTEM_ERROR),
@@ -231,7 +224,7 @@ seccomp_test_sighandler(int signum, siginfo_t *info, void *cxt pg_attribute_unus
 		die(1, DIE_PREFIX "bad signal number\n");

 	/* TODO: maybe somehow extract the hardcoded syscall number */
-	if (info->si_syscall != SCMP_SYS(openat))
+	if (info->si_syscall != SCMP_SYS(open))
 		die(1, DIE_PREFIX "bad syscall number\n");

 #undef DIE_PREFIX
--- a/poetry.lock
+++ b/poetry.lock
@@ -1,4 +1,4 @@
-# This file is automatically @generated by Poetry 1.4.1 and should not be changed by hand.
+# This file is automatically @generated by Poetry 1.4.0 and should not be changed by hand.

 [[package]]
 name = "aiohttp"
@@ -79,35 +79,37 @@ sa = ["sqlalchemy[postgresql-psycopg2binary] (>=1.3,<1.5)"]

 [[package]]
 name = "allure-pytest"
-version = "2.13.1"
+version = "2.10.0"
 description = "Allure pytest integration"
 category = "main"
 optional = false
 python-versions = "*"
 files = [
-    {file = "allure-pytest-2.13.1.tar.gz", hash = "sha256:68d69456eeb65af4061ec06a80bc941163b0616e8216554d36b070a6bf070e08"},
-    {file = "allure_pytest-2.13.1-py3-none-any.whl", hash = "sha256:a8de2fc3b3effe2d8f98801646920de3f055b779710f4c806dbee7c613c24633"},
+    {file = "allure-pytest-2.10.0.tar.gz", hash = "sha256:3b2ab67629f4cbd8617abd817d2b22292c6eb7efd5584f992d1af8143aea6ee7"},
+    {file = "allure_pytest-2.10.0-py3-none-any.whl", hash = "sha256:08274096594758447db54c3b2c382526ee04f1fe12119cdaee92d2d93c84b530"},
 ]

 [package.dependencies]
-allure-python-commons = "2.13.1"
+allure-python-commons = "2.10.0"
 pytest = ">=4.5.0"
+six = ">=1.9.0"

 [[package]]
 name = "allure-python-commons"
-version = "2.13.1"
+version = "2.10.0"
 description = "Common module for integrate allure with python-based frameworks"
 category = "main"
 optional = false
-python-versions = ">=3.6"
+python-versions = ">=3.5"
 files = [
-    {file = "allure-python-commons-2.13.1.tar.gz", hash = "sha256:3fc13e1da8ebb23f9ab5c9c72ad04595023cdd5078dbb8604939997faebed5cb"},
-    {file = "allure_python_commons-2.13.1-py3-none-any.whl", hash = "sha256:d08e04867bddf44fef55def3d67f4bc25af58a1bf9fcffcf4ec3331f7f2ef0d0"},
+    {file = "allure-python-commons-2.10.0.tar.gz", hash = "sha256:d4d31344b0f0037a4a11e16b91b28cf0eeb23ffa0e50c27fcfc6aabe72212d3c"},
+    {file = "allure_python_commons-2.10.0-py3-none-any.whl", hash = "sha256:2a717e8ca8d296bf89cd57f38fc3c21893bd7ea8cd02a6ae5420e6d1a6eda5d0"},
 ]

 [package.dependencies]
 attrs = ">=16.0.0"
 pluggy = ">=0.4.0"
+six = ">=1.9.0"

 [[package]]
 name = "async-timeout"
@@ -1930,22 +1932,6 @@ pytest = [
    {version = ">=6.2.4", markers = "python_version >= \"3.10\""},
 ]

-[[package]]
-name = "pytest-rerunfailures"
-version = "11.1.2"
-description = "pytest plugin to re-run tests to eliminate flaky failures"
-category = "main"
-optional = false
-python-versions = ">=3.7"
-files = [
-    {file = "pytest-rerunfailures-11.1.2.tar.gz", hash = "sha256:55611661e873f1cafa384c82f08d07883954f4b76435f4b8a5b470c1954573de"},
-    {file = "pytest_rerunfailures-11.1.2-py3-none-any.whl", hash = "sha256:d21fe2e46d9774f8ad95f1aa799544ae95cac3a223477af94aa985adfae92b7e"},
-]
-
-[package.dependencies]
-packaging = ">=17.1"
-pytest = ">=5.3"
-
 [[package]]
 name = "pytest-timeout"
 version = "2.1.0"
@@ -2611,4 +2597,4 @@ testing = ["func-timeout", "jaraco.itertools", "pytest (>=6)", "pytest-black (>=
 [metadata]
 lock-version = "2.0"
 python-versions = "^3.9"
-content-hash = "b689ffd6eae32b966f1744b5ac3343fe0dd26b31ee1f50e13daf5045ee0623e1"
+content-hash = "2515a9320c2960076012fbc036fb33c4f6a23515c8d143785931dc18c6722d91"
--- a/proxy/Cargo.toml
+++ b/proxy/Cargo.toml
@@ -64,7 +64,6 @@ webpki-roots.workspace = true
 x509-parser.workspace = true

 workspace_hack.workspace = true
-tokio-util.workspace = true

 [dev-dependencies]
 rcgen.workspace = true
--- a/proxy/src/auth/backend.rs
+++ b/proxy/src/auth/backend.rs
@@ -140,7 +140,7 @@ async fn auth_quirks(

 impl BackendType<'_, ClientCredentials<'_>> {
    /// Authenticate the client via the requested backend, possibly using credentials.
-    #[tracing::instrument(fields(allow_cleartext = allow_cleartext), skip_all)]
+    #[tracing::instrument(fields(allow_cleartext), skip_all)]
    pub async fn authenticate(
        &mut self,
        extra: &ConsoleReqExtra<'_>,
--- a/proxy/src/auth/backend/hacks.rs
+++ b/proxy/src/auth/backend/hacks.rs
@@ -53,7 +53,7 @@ pub async fn password_hack(
        .await?;

    info!(project = &payload.project, "received missing parameter");
-    creds.project = Some(payload.project);
+    creds.project = Some(payload.project.into());

    let mut node = api.wake_compute(extra, creds).await?;
    node.config.password(payload.password);
--- a/proxy/src/auth/credentials.rs
+++ b/proxy/src/auth/credentials.rs
@@ -2,7 +2,7 @@

 use crate::error::UserFacingError;
 use pq_proto::StartupMessageParams;
-use std::collections::HashSet;
+use std::borrow::Cow;
 use thiserror::Error;
 use tracing::info;

@@ -19,10 +19,11 @@ pub enum ClientCredsParseError {
    InconsistentProjectNames { domain: String, option: String },

    #[error(
-        "Common name inferred from SNI ('{}') is not known",
-        .cn,
+        "SNI ('{}') inconsistently formatted with respect to common name ('{}'). \
+         SNI should be formatted as '<project-name>.{}'.",
+        .sni, .cn, .cn,
    )]
-    UnknownCommonName { cn: String },
+    InconsistentSni { sni: String, cn: String },

    #[error("Project name ('{0}') must contain only alphanumeric characters and hyphen.")]
    MalformedProjectName(String),
@@ -36,7 +37,7 @@ impl UserFacingError for ClientCredsParseError {}
 pub struct ClientCredentials<'a> {
    pub user: &'a str,
    // TODO: this is a severe misnomer! We should think of a new name ASAP.
-    pub project: Option<String>,
+    pub project: Option<Cow<'a, str>>,
 }

 impl ClientCredentials<'_> {
@@ -50,7 +51,7 @@ impl<'a> ClientCredentials<'a> {
    pub fn parse(
        params: &'a StartupMessageParams,
        sni: Option<&str>,
-        common_names: Option<HashSet<String>>,
+        common_name: Option<&str>,
    ) -> Result<Self, ClientCredsParseError> {
        use ClientCredsParseError::*;

@@ -59,43 +60,37 @@ impl<'a> ClientCredentials<'a> {
        let user = get_param("user")?;

        // Project name might be passed via PG's command-line options.
-        let project_option = params
-            .options_raw()
-            .and_then(|mut options| options.find_map(|opt| opt.strip_prefix("project=")))
-            .map(|name| name.to_string());
+        let project_option = params.options_raw().and_then(|mut options| {
+            options
+                .find_map(|opt| opt.strip_prefix("project="))
+                .map(Cow::Borrowed)
+        });

-        let project_from_domain = if let Some(sni_str) = sni {
-            if let Some(cn) = common_names {
-                let common_name_from_sni = sni_str.split_once('.').map(|(_, domain)| domain);
-
-                let project = common_name_from_sni
-                    .and_then(|domain| {
-                        if cn.contains(domain) {
-                            subdomain_from_sni(sni_str, domain)
-                        } else {
-                            None
-                        }
+        // Alternative project name is in fact a subdomain from SNI.
+        // NOTE: we do not consider SNI if `common_name` is missing.
+        let project_domain = sni
+            .zip(common_name)
+            .map(|(sni, cn)| {
+                subdomain_from_sni(sni, cn)
+                    .ok_or_else(|| InconsistentSni {
+                        sni: sni.into(),
+                        cn: cn.into(),
                    })
-                    .ok_or_else(|| UnknownCommonName {
-                        cn: common_name_from_sni.unwrap_or("").into(),
-                    })?;
+                    .map(Cow::<'static, str>::Owned)
+            })
+            .transpose()?;

-                Some(project)
-            } else {
-                None
-            }
-        } else {
-            None
-        };
-
-        let project = match (project_option, project_from_domain) {
+        let project = match (project_option, project_domain) {
            // Invariant: if we have both project name variants, they should match.
            (Some(option), Some(domain)) if option != domain => {
-                Some(Err(InconsistentProjectNames { domain, option }))
+                Some(Err(InconsistentProjectNames {
+                    domain: domain.into(),
+                    option: option.into(),
+                }))
            }
            // Invariant: project name may not contain certain characters.
            (a, b) => a.or(b).map(|name| match project_name_valid(&name) {
-                false => Err(MalformedProjectName(name)),
+                false => Err(MalformedProjectName(name.into())),
                true => Ok(name),
            }),
        }
@@ -154,9 +149,9 @@ mod tests {
        let options = StartupMessageParams::new([("user", "john_doe")]);

        let sni = Some("foo.localhost");
-        let common_names = Some(["localhost".into()].into());
+        let common_name = Some("localhost");

-        let creds = ClientCredentials::parse(&options, sni, common_names)?;
+        let creds = ClientCredentials::parse(&options, sni, common_name)?;
        assert_eq!(creds.user, "john_doe");
        assert_eq!(creds.project.as_deref(), Some("foo"));

@@ -182,41 +177,24 @@ mod tests {
        let options = StartupMessageParams::new([("user", "john_doe"), ("options", "project=baz")]);

        let sni = Some("baz.localhost");
-        let common_names = Some(["localhost".into()].into());
+        let common_name = Some("localhost");

-        let creds = ClientCredentials::parse(&options, sni, common_names)?;
+        let creds = ClientCredentials::parse(&options, sni, common_name)?;
        assert_eq!(creds.user, "john_doe");
        assert_eq!(creds.project.as_deref(), Some("baz"));

        Ok(())
    }

-    #[test]
-    fn parse_multi_common_names() -> anyhow::Result<()> {
-        let options = StartupMessageParams::new([("user", "john_doe")]);
-
-        let common_names = Some(["a.com".into(), "b.com".into()].into());
-        let sni = Some("p1.a.com");
-        let creds = ClientCredentials::parse(&options, sni, common_names)?;
-        assert_eq!(creds.project.as_deref(), Some("p1"));
-
-        let common_names = Some(["a.com".into(), "b.com".into()].into());
-        let sni = Some("p1.b.com");
-        let creds = ClientCredentials::parse(&options, sni, common_names)?;
-        assert_eq!(creds.project.as_deref(), Some("p1"));
-
-        Ok(())
-    }
-
    #[test]
    fn parse_projects_different() {
        let options =
            StartupMessageParams::new([("user", "john_doe"), ("options", "project=first")]);

        let sni = Some("second.localhost");
-        let common_names = Some(["localhost".into()].into());
+        let common_name = Some("localhost");

-        let err = ClientCredentials::parse(&options, sni, common_names).expect_err("should fail");
+        let err = ClientCredentials::parse(&options, sni, common_name).expect_err("should fail");
        match err {
            InconsistentProjectNames { domain, option } => {
                assert_eq!(option, "first");
@@ -231,12 +209,13 @@ mod tests {
        let options = StartupMessageParams::new([("user", "john_doe")]);

        let sni = Some("project.localhost");
-        let common_names = Some(["example.com".into()].into());
+        let common_name = Some("example.com");

-        let err = ClientCredentials::parse(&options, sni, common_names).expect_err("should fail");
+        let err = ClientCredentials::parse(&options, sni, common_name).expect_err("should fail");
        match err {
-            UnknownCommonName { cn } => {
-                assert_eq!(cn, "localhost");
+            InconsistentSni { sni, cn } => {
+                assert_eq!(sni, "project.localhost");
+                assert_eq!(cn, "example.com");
            }
            _ => panic!("bad error: {err:?}"),
        }
--- a/proxy/src/config.rs
+++ b/proxy/src/config.rs
@@ -1,12 +1,6 @@
 use crate::auth;
-use anyhow::{bail, ensure, Context, Ok};
-use rustls::sign;
-use std::{
-    collections::{HashMap, HashSet},
-    str::FromStr,
-    sync::Arc,
-    time::Duration,
-};
+use anyhow::{bail, ensure, Context};
+use std::{str::FromStr, sync::Arc, time::Duration};

 pub struct ProxyConfig {
    pub tls_config: Option<TlsConfig>,
@@ -22,7 +16,7 @@ pub struct MetricCollectionConfig {

 pub struct TlsConfig {
    pub config: Arc<rustls::ServerConfig>,
-    pub common_names: Option<HashSet<String>>,
+    pub common_name: Option<String>,
 }

 impl TlsConfig {
@@ -32,37 +26,28 @@ impl TlsConfig {
 }

 /// Configure TLS for the main endpoint.
-pub fn configure_tls(
-    key_path: &str,
-    cert_path: &str,
-    certs_dir: Option<&String>,
-) -> anyhow::Result<TlsConfig> {
-    let mut cert_resolver = CertResolver::new();
+pub fn configure_tls(key_path: &str, cert_path: &str) -> anyhow::Result<TlsConfig> {
+    let key = {
+        let key_bytes = std::fs::read(key_path).context("TLS key file")?;
+        let mut keys = rustls_pemfile::pkcs8_private_keys(&mut &key_bytes[..])
+            .context(format!("Failed to read TLS keys at '{key_path}'"))?;

-    // add default certificate
-    cert_resolver.add_cert(key_path, cert_path, true)?;
+        ensure!(keys.len() == 1, "keys.len() = {} (should be 1)", keys.len());
+        keys.pop().map(rustls::PrivateKey).unwrap()
+    };

-    // add extra certificates
-    if let Some(certs_dir) = certs_dir {
-        for entry in std::fs::read_dir(certs_dir)? {
-            let entry = entry?;
-            let path = entry.path();
-            if path.is_dir() {
-                // file names aligned with default cert-manager names
-                let key_path = path.join("tls.key");
-                let cert_path = path.join("tls.crt");
-                if key_path.exists() && cert_path.exists() {
-                    cert_resolver.add_cert(
-                        &key_path.to_string_lossy(),
-                        &cert_path.to_string_lossy(),
-                        false,
-                    )?;
-                }
-            }
-        }
-    }
+    let cert_chain_bytes = std::fs::read(cert_path)
+        .context(format!("Failed to read TLS cert file at '{cert_path}.'"))?;

-    let common_names = cert_resolver.get_common_names();
+    let cert_chain = {
+        rustls_pemfile::certs(&mut &cert_chain_bytes[..])
+            .context(format!(
+                "Failed to read TLS certificate chain from bytes from file at '{cert_path}'."
+            ))?
+            .into_iter()
+            .map(rustls::Certificate)
+            .collect()
+    };

    let config = rustls::ServerConfig::builder()
        .with_safe_default_cipher_suites()
@@ -70,136 +55,27 @@ pub fn configure_tls(
        // allow TLS 1.2 to be compatible with older client libraries
        .with_protocol_versions(&[&rustls::version::TLS13, &rustls::version::TLS12])?
        .with_no_client_auth()
-        .with_cert_resolver(Arc::new(cert_resolver))
+        .with_single_cert(cert_chain, key)?
        .into();

+    // determine common name from tls-cert (-c server.crt param).
+    // used in asserting project name formatting invariant.
+    let common_name = {
+        let pem = x509_parser::pem::parse_x509_pem(&cert_chain_bytes)
+            .context(format!(
+                "Failed to parse PEM object from bytes from file at '{cert_path}'."
+            ))?
+            .1;
+        let common_name = pem.parse_x509()?.subject().to_string();
+        common_name.strip_prefix("CN=*.").map(|s| s.to_string())
+    };
+
    Ok(TlsConfig {
        config,
-        common_names: Some(common_names),
+        common_name,
    })
 }

-struct CertResolver {
-    certs: HashMap<String, Arc<rustls::sign::CertifiedKey>>,
-    default: Option<Arc<rustls::sign::CertifiedKey>>,
-}
-
-impl CertResolver {
-    fn new() -> Self {
-        Self {
-            certs: HashMap::new(),
-            default: None,
-        }
-    }
-
-    fn add_cert(
-        &mut self,
-        key_path: &str,
-        cert_path: &str,
-        is_default: bool,
-    ) -> anyhow::Result<()> {
-        let priv_key = {
-            let key_bytes = std::fs::read(key_path).context("TLS key file")?;
-            let mut keys = rustls_pemfile::pkcs8_private_keys(&mut &key_bytes[..])
-                .context(format!("Failed to read TLS keys at '{key_path}'"))?;
-
-            ensure!(keys.len() == 1, "keys.len() = {} (should be 1)", keys.len());
-            keys.pop().map(rustls::PrivateKey).unwrap()
-        };
-
-        let key = sign::any_supported_type(&priv_key).context("invalid private key")?;
-
-        let cert_chain_bytes = std::fs::read(cert_path)
-            .context(format!("Failed to read TLS cert file at '{cert_path}.'"))?;
-
-        let cert_chain = {
-            rustls_pemfile::certs(&mut &cert_chain_bytes[..])
-                .context(format!(
-                    "Failed to read TLS certificate chain from bytes from file at '{cert_path}'."
-                ))?
-                .into_iter()
-                .map(rustls::Certificate)
-                .collect()
-        };
-
-        let common_name = {
-            let pem = x509_parser::pem::parse_x509_pem(&cert_chain_bytes)
-                .context(format!(
-                    "Failed to parse PEM object from bytes from file at '{cert_path}'."
-                ))?
-                .1;
-            let common_name = pem.parse_x509()?.subject().to_string();
-
-            // We only use non-wildcard certificates in link proxy so it seems okay to treat them the same as
-            // wildcard ones as we don't use SNI there. That treatment only affects certificate selection, so
-            // verify-full will still check wildcard match. Old coding here just ignored non-wildcard common names
-            // and passed None instead, which blows up number of cases downstream code should handle. Proper coding
-            // here should better avoid Option for common_names, and do wildcard-based certificate selection instead
-            // of cutting off '*.' parts.
-            if common_name.starts_with("CN=*.") {
-                common_name.strip_prefix("CN=*.").map(|s| s.to_string())
-            } else {
-                common_name.strip_prefix("CN=").map(|s| s.to_string())
-            }
-        }
-        .context(format!(
-            "Failed to parse common name from certificate at '{cert_path}'."
-        ))?;
-
-        let cert = Arc::new(rustls::sign::CertifiedKey::new(cert_chain, key));
-
-        if is_default {
-            self.default = Some(cert.clone());
-        }
-
-        self.certs.insert(common_name, cert);
-
-        Ok(())
-    }
-
-    fn get_common_names(&self) -> HashSet<String> {
-        self.certs.keys().map(|s| s.to_string()).collect()
-    }
-}
-
-impl rustls::server::ResolvesServerCert for CertResolver {
-    fn resolve(
-        &self,
-        _client_hello: rustls::server::ClientHello,
-    ) -> Option<Arc<rustls::sign::CertifiedKey>> {
-        // loop here and cut off more and more subdomains until we find
-        // a match to get a proper wildcard support. OTOH, we now do not
-        // use nested domains, so keep this simple for now.
-        //
-        // With the current coding foo.com will match *.foo.com and that
-        // repeats behavior of the old code.
-        if let Some(mut sni_name) = _client_hello.server_name() {
-            loop {
-                if let Some(cert) = self.certs.get(sni_name) {
-                    return Some(cert.clone());
-                }
-                if let Some((_, rest)) = sni_name.split_once('.') {
-                    sni_name = rest;
-                } else {
-                    return None;
-                }
-            }
-        } else {
-            // No SNI, use the default certificate, otherwise we can't get to
-            // options parameter which can be used to set endpoint name too.
-            // That means that non-SNI flow will not work for CNAME domains in
-            // verify-full mode.
-            //
-            // If that will be a problem we can:
-            //
-            // a) Instead of multi-cert approach use single cert with extra
-            //    domains listed in Subject Alternative Name (SAN).
-            // b) Deploy separate proxy instances for extra domains.
-            self.default.as_ref().cloned()
-        }
-    }
-}
-
 /// Helper for cmdline cache options parsing.
 pub struct CacheOptions {
    /// Max number of entries.
--- a/Show More
+++ b/Show More