greptimedb/.github/workflows/nightly-jsonbench.yaml

name: Nightly JSONBench

on:
  workflow_run:
    workflows: [ "GreptimeDB Nightly Build" ]
    types: [ completed ]
  workflow_dispatch:
    inputs:
      run_id:
        description: The nightly build workflow run id to download GreptimeDB artifacts from
        required: true
        type: string

permissions:
  actions: read
  contents: read

concurrency:
  group: ${{ github.workflow }}-${{ github.head_ref || github.run_id }}
  cancel-in-progress: true

jobs:
  resolve-artifact:
    name: Resolve GreptimeDB nightly artifact
    if: ${{ github.repository == 'GreptimeTeam/greptimedb' && (github.event_name == 'workflow_dispatch' || github.event.workflow_run.conclusion == 'success') }}
    runs-on: ubuntu-latest
    outputs:
      artifact-name: ${{ steps.find-artifact.outputs.artifact-name }}
      run-id: ${{ steps.resolve-run-id.outputs.run-id }}
    steps:
      - name: Resolve nightly build run id
        id: resolve-run-id
        shell: bash
        env:
          EVENT_NAME: ${{ github.event_name }}
          WORKFLOW_RUN_ID: ${{ github.event.workflow_run.id }}
          INPUT_RUN_ID: ${{ inputs.run_id }}
        run: |
          set -euo pipefail

          if [[ "${EVENT_NAME}" == "workflow_dispatch" ]]; then
            run_id="${INPUT_RUN_ID}"
          else
            run_id="${WORKFLOW_RUN_ID}"
          fi

          if [[ ! "${run_id}" =~ ^[0-9]+$ ]]; then
            echo "Invalid workflow run id: ${run_id}"
            exit 1
          fi

          echo "run-id=${run_id}" >> "${GITHUB_OUTPUT}"

      - name: Find GreptimeDB nightly artifact
        id: find-artifact
        shell: bash
        env:
          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
          RUN_ID: ${{ steps.resolve-run-id.outputs.run-id }}
        run: |
          set -euo pipefail

          artifact_name=$(gh api "repos/${GITHUB_REPOSITORY}/actions/runs/${RUN_ID}/artifacts" --paginate \
            --jq '.artifacts[] | select(.name | test("^greptime-linux-arm64-nightly-[0-9]{8}-[0-9a-f]+$")) | .name' \
            | head -n 1)

          if [[ -z "${artifact_name}" ]]; then
            echo "Cannot find linux arm64 nightly artifact in workflow run ${RUN_ID}."
            exit 1
          fi

          echo "Download GreptimeDB artifact: ${artifact_name}"
          echo "artifact-name=${artifact_name}" >> "${GITHUB_OUTPUT}"

  allocate-runner:
    name: Allocate runner
    if: ${{ github.repository == 'GreptimeTeam/greptimedb' && (github.event_name == 'workflow_dispatch' || github.event.workflow_run.conclusion == 'success') }}
    needs: [ resolve-artifact ]
    runs-on: ubuntu-latest
    outputs:
      linux-arm64-runner: ${{ steps.start-linux-arm64-runner.outputs.label }}

      # The following EC2 resource id will be used for resource releasing.
      linux-arm64-ec2-runner-label: ${{ steps.start-linux-arm64-runner.outputs.label }}
      linux-arm64-ec2-runner-instance-id: ${{ steps.start-linux-arm64-runner.outputs.ec2-instance-id }}
    steps:
      - name: Checkout
        uses: actions/checkout@v4
        with:
          fetch-depth: 0
          persist-credentials: false

      - name: Allocate Linux ARM64 runner
        uses: ./.github/actions/start-runner
        id: start-linux-arm64-runner
        with:
          runner: ${{ vars.DEFAULT_ARM64_RUNNER }}
          aws-access-key-id: ${{ secrets.AWS_ACCESS_KEY_ID }}
          aws-secret-access-key: ${{ secrets.AWS_SECRET_ACCESS_KEY }}
          aws-region: ${{ vars.EC2_RUNNER_REGION }}
          github-token: ${{ secrets.GH_PERSONAL_ACCESS_TOKEN }}
          image-id: ${{ vars.EC2_RUNNER_LINUX_ARM64_IMAGE_ID }}
          security-group-id: ${{ vars.EC2_RUNNER_SECURITY_GROUP_ID }}
          subnet-id: ${{ vars.EC2_RUNNER_SUBNET_ID }}

  jsonbench:
    name: Run JSONBench
    if: ${{ github.repository == 'GreptimeTeam/greptimedb' && (github.event_name == 'workflow_dispatch' || github.event.workflow_run.conclusion == 'success') }}
    needs: [ resolve-artifact, allocate-runner ]
    runs-on: ${{ needs.allocate-runner.outputs.linux-arm64-runner }}
    timeout-minutes: 120
    env:
      JSONBENCH_OUTPUT_PREFIX: _linux-arm64
    steps:
      - name: Download GreptimeDB nightly artifact
        uses: actions/download-artifact@v4
        with:
          name: ${{ needs.resolve-artifact.outputs.artifact-name }}
          path: greptimedb-artifact
          github-token: ${{ secrets.GITHUB_TOKEN }}
          run-id: ${{ needs.resolve-artifact.outputs.run-id }}

      - name: Prepare GreptimeDB binary
        shell: bash
        run: |
          set -euo pipefail

          tar -xzf "greptimedb-artifact/${{ needs.resolve-artifact.outputs.artifact-name }}.tar.gz"
          cp "${{ needs.resolve-artifact.outputs.artifact-name }}/greptime" ./greptime
          chmod +x ./greptime
          rm -rf greptimedb-artifact "${{ needs.resolve-artifact.outputs.artifact-name }}"

      - name: Run JSONBench
        env:
          # TODO(LFC): Change to "3" (100m) when JSON2 ingestion performance is optimized.
          JSONBENCH_DATASET: 2
        shell: bash
        run: |
          set -euo pipefail

          export JSONBENCH_DATA_DIR="/root/data/bluesky"
          echo "Use JSONBench data directory ${JSONBENCH_DATA_DIR}"

          echo "Cloning JSONBench"
          git clone --branch greptimedb-new-json --depth 1 https://github.com/GreptimeTeam/JSONBench.git JSONBench

          echo "Downloading JSONBench dataset choice ${JSONBENCH_DATASET} to ${JSONBENCH_DATA_DIR}"
          mkdir -p "${JSONBENCH_DATA_DIR}"
          printf "${JSONBENCH_DATASET}\n" | ./JSONBench/download_data.sh
          downloaded_files=$(find "${JSONBENCH_DATA_DIR}" -type f | wc -l)
          echo "Downloaded JSONBench dataset files: ${downloaded_files}"

          export GREPTIMEDB_STANDALONE__WAL__DIR=greptimedb_data/wal
          export GREPTIMEDB_STANDALONE__STORAGE__DATA_HOME=greptimedb_data
          export GREPTIMEDB_STANDALONE__LOGGING__DIR=greptimedb_data/logs
          export GREPTIMEDB_STANDALONE__LOGGING__APPEND_STDOUT=false
          export GREPTIMEDB_STANDALONE__HTTP__BODY_LIMIT=1GB
          export GREPTIMEDB_STANDALONE__HTTP__TIMEOUT=500s

          echo "Starting GreptimeDB standalone"
          ./greptime standalone start > greptimedb.log 2>&1 &
          greptime_pid=$!
          trap 'kill "${greptime_pid}" 2>/dev/null || true' EXIT

          echo "Waiting for GreptimeDB health check"
          until curl -s --fail -o /dev/null http://localhost:4000/health; do
            if ! kill -0 "${greptime_pid}" 2>/dev/null; then
              cat greptimedb.log
              exit 1
            fi
            sleep 1
          done
          echo "GreptimeDB is ready"

          cp ./greptime JSONBench/greptimedb/greptime

          cd JSONBench/greptimedb
          echo "Running JSONBench main.sh with dataset choice ${JSONBENCH_DATASET} and install=false"
          ./main.sh ${JSONBENCH_DATASET} "${JSONBENCH_DATA_DIR}" success.log error.log "${JSONBENCH_OUTPUT_PREFIX}" false
          echo "JSONBench finished"

      - name: Upload JSONBench results
        if: always()
        uses: actions/upload-artifact@v4
        with:
          name: jsonbench-results
          path: |
            ./greptimedb.log
            ./JSONBench/greptimedb/*.log
            ./JSONBench/greptimedb/*.total_size
            ./JSONBench/greptimedb/*.data_size
            ./JSONBench/greptimedb/*.index_size
            ./JSONBench/greptimedb/*.count
            ./JSONBench/greptimedb/*.results_runtime
            ./JSONBench/greptimedb/*.query_results
          if-no-files-found: ignore
          retention-days: 7

  stop-linux-arm64-runner:
    name: Stop Linux ARM64 runner
    # It's always run as the last job in the workflow to make sure that the runner is released.
    if: ${{ always() && needs.allocate-runner.outputs.linux-arm64-ec2-runner-instance-id != '' }}
    runs-on: ubuntu-latest
    needs: [
      allocate-runner,
      jsonbench,
    ]
    steps:
      - name: Checkout
        uses: actions/checkout@v4
        with:
          fetch-depth: 0
          persist-credentials: false

      - name: Stop Linux ARM64 runner
        uses: ./.github/actions/stop-runner
        with:
          label: ${{ needs.allocate-runner.outputs.linux-arm64-ec2-runner-label }}
          ec2-instance-id: ${{ needs.allocate-runner.outputs.linux-arm64-ec2-runner-instance-id }}
          aws-access-key-id: ${{ secrets.AWS_ACCESS_KEY_ID }}
          aws-secret-access-key: ${{ secrets.AWS_SECRET_ACCESS_KEY }}
          aws-region: ${{ vars.EC2_RUNNER_REGION }}
          github-token: ${{ secrets.GH_PERSONAL_ACCESS_TOKEN }}