Jinwoo HongandClaude 76f9978880 fix(relay): judge the reserve watch's 503 rule on non-drain 503s, not a host's own re-dials (#30487) ...

* fix(relay): judge the reserve watch's 503 rule on non-drain 503s, not a host's own re-dials

The watch counted every director 503, and on 2026-10-18 ~98% were host-rate-limited re-dials that
grow with traffic, so it flipped production-gce-c7 back twice with no fault. It now uses the roll
ladder's non-drain measure (request_count 503s less drain-return deferrals and host-rate-limited /
host-in-flight re-dials from the director's runtime metrics) for both the background and the live
samples. A judged minute no runtime-metrics line covers is unverifiable, not zero.

Co-Authored-By: Claude <noreply@anthropic.com>

* fix(relay): state the reserve watch's metrics coverage honestly and share the metrics read limit

Co-Authored-By: Claude <noreply@anthropic.com>

---------

Co-authored-by: Claude <noreply@anthropic.com>

2026-10-18 22:40:33 -04:00

..

adhoc-mac-build.yml

fix(macos): preserve terminal permissions across app updates with a small signed helper (#27021)

2026-10-11 14:54:36 -04:00

agent-state-rules-publish.yml

feat: live updates for agent state rules (#24387)

2026-10-02 14:59:24 -04:00

ci-cache-warmup-platform-jobs.yml

ci: move PR analysis and Linux unit tests to Blacksmith (#29362)

2026-10-15 21:07:49 -07:00

ci-cache-warmup-platforms.yml

ci: skip checks for draft pull requests [skip ci] (#29484)

2026-10-15 23:18:13 -07:00

ci-cache-warmup.yml

ci: skip checks for draft pull requests [skip ci] (#29484)

2026-10-15 23:18:13 -07:00

ci-closed-pr-caches.yml

ci: cancel obsolete PR checks even when new checks are skipped (#29364)

2026-10-15 20:12:36 -07:00

ci-pnpm-verification-pilot.yml

ci: skip checks for draft pull requests [skip ci] (#29484)

2026-10-15 23:18:13 -07:00

ci-runner-demand.yml

ci: reduce repeated runner work and validate affected-test selection (#23540)

2026-09-27 23:25:20 -07:00

ci-xterm-cache.yml

ci: skip unrelated installs and share xterm build dependencies (#23607)

2026-09-28 03:46:22 -07:00

cloud-bootstrap-relay-staging-capacity.yml

ci(relay): use the runner's preinstalled gcloud instead of downloading the SDK (#28392)

2026-10-13 16:24:53 -04:00

cloud-deploy-relay-asia-topology.yml

ci(relay): use the runner's preinstalled gcloud instead of downloading the SDK (#28392)

2026-10-13 16:24:53 -04:00

cloud-deploy-relay-fence-broker.yml

ci(relay): use the runner's preinstalled gcloud instead of downloading the SDK (#28392)

2026-10-13 16:24:53 -04:00

cloud-deploy-relay-production-capacity-job.yml

fix(relay): end the restart-safe quiet at the drain's pace window, not a window after it (#28677)

2026-10-14 15:35:31 -04:00

cloud-deploy-relay-production-capacity.yml

ci(relay): use the runner's preinstalled gcloud instead of downloading the SDK (#28392)

2026-10-13 16:24:53 -04:00

cloud-deploy-relay-production-director.yml

ci(relay): use the runner's preinstalled gcloud instead of downloading the SDK (#28392)

2026-10-13 16:24:53 -04:00

cloud-deploy-relay-production-multi-target.yml

ci(relay): use the runner's preinstalled gcloud instead of downloading the SDK (#28392)

2026-10-13 16:24:53 -04:00

cloud-deploy-relay-production-same-cap-job.yml

fix(relay): end the restart-safe quiet at the drain's pace window, not a window after it (#28677)

2026-10-14 15:35:31 -04:00

cloud-deploy-relay-production-same-cap.yml

ci(relay): use the runner's preinstalled gcloud instead of downloading the SDK (#28392)

2026-10-13 16:24:53 -04:00

cloud-deploy-relay-production.yml

ci(relay): use the runner's preinstalled gcloud instead of downloading the SDK (#28392)

2026-10-13 16:24:53 -04:00

cloud-deploy-relay-staging-gce-candidate.yml

ci(relay): use the runner's preinstalled gcloud instead of downloading the SDK (#28392)

2026-10-13 16:24:53 -04:00

cloud-deploy-relay-staging.yml

ci(relay): use the runner's preinstalled gcloud instead of downloading the SDK (#28392)

2026-10-13 16:24:53 -04:00

cloud-monitor-relay-clock-skew.yml

fix(relay): check clock skew on every cell production.tfvars declares (#28786)

2026-10-15 19:12:48 -07:00

cloud-monitor-relay-production-job.yml

ci(relay): use the runner's preinstalled gcloud instead of downloading the SDK (#28392)

2026-10-13 16:24:53 -04:00

cloud-monitor-relay-production.yml

chore(cloud): add the relay fence broker, ops console, Terraform root, scripts, and 24 cloud-* workflows (#18413)

2026-09-03 06:55:14 -04:00

cloud-operate-relay-asia-admission.yml

ci(relay): use the runner's preinstalled gcloud instead of downloading the SDK (#28392)

2026-10-13 16:24:53 -04:00

cloud-operate-relay-production-cell-flags.yml

ci(relay): use the runner's preinstalled gcloud instead of downloading the SDK (#28392)

2026-10-13 16:24:53 -04:00

cloud-operate-relay-production-rehome-job.yml

ci(relay): use the runner's preinstalled gcloud instead of downloading the SDK (#28392)

2026-10-13 16:24:53 -04:00

cloud-operate-relay-production-rehome.yml

fix(relay): rehome hosts to their preferred region in either direction (#19241)

2026-09-07 04:40:37 -04:00

cloud-operate-relay-production-reserve-watch.yml

fix(relay): judge the reserve watch's 503 rule on non-drain 503s, not a host's own re-dials (#30487)

2026-10-18 22:40:33 -04:00

cloud-operate-relay-staging-cell-flags.yml

ci(relay): use the runner's preinstalled gcloud instead of downloading the SDK (#28392)

2026-10-13 16:24:53 -04:00

cloud-power-relay-staging.yml

ci(relay): use the runner's preinstalled gcloud instead of downloading the SDK (#28392)

2026-10-13 16:24:53 -04:00

cloud-prove-relay-asia-staging.yml

ci(relay-staging): split the staging load across two hosted runners (#27402)

2026-10-11 02:48:37 -04:00

cloud-prove-relay-staging-capacity.yml

fix(relay): end the restart-safe quiet at the drain's pace window, not a window after it (#28677)

2026-10-14 15:35:31 -04:00

cloud-publish-relay-production.yml

ci(relay): use the runner's preinstalled gcloud instead of downloading the SDK (#28392)

2026-10-13 16:24:53 -04:00

cloud-push-deploy.yml

ci(relay): use the runner's preinstalled gcloud instead of downloading the SDK (#28392)

2026-10-13 16:24:53 -04:00

cloud-recover-relay-staging-c4-image.yml

ci(relay): use the runner's preinstalled gcloud instead of downloading the SDK (#28392)

2026-10-13 16:24:53 -04:00

cloud-requeue-relay-staging-c4-recovery.yml

chore(cloud): add the relay fence broker, ops console, Terraform root, scripts, and 24 cloud-* workflows (#18413)

2026-09-03 06:55:14 -04:00

cloud-verify.yml

ci: skip checks for draft pull requests [skip ci] (#29484)

2026-10-15 23:18:13 -07:00

computer-e2e.yml

[CI] Reduce test imports, clock waits, and headless cleanup costs (#29312)

2026-10-15 23:31:35 -07:00

daemon-relocation-spike.yml

feat(mobile): one build-time switch picks native or OTA, default native (OTA phase E1) (#22193)

2026-09-22 06:18:33 -04:00

daily-mac-build.yml

fix(macos): preserve terminal permissions across app updates with a small signed helper (#27021)

2026-10-11 14:54:36 -04:00

dev-channel-win-build.yml

ci: drop the relay Windows process-tree job from desktop packaging (#27412)

2026-10-11 01:32:14 -07:00

docs.yml

ci: skip checks for draft pull requests [skip ci] (#29484)

2026-10-15 23:18:13 -07:00

e2e.yml

feat(orcad): ship the mobile web page so phones paired to orcad get it (#29882)

2026-10-16 19:29:41 -07:00

git-command-termination-runtime.yml

ci: skip checks for draft pull requests [skip ci] (#29484)

2026-10-15 23:18:13 -07:00

golden-e2e-experiment.yml

fix(ci): preserve focused Playwright file selection (#23270)

2026-09-26 15:10:35 -07:00

homebrew-bump.yml

chore(ci): stop auto-filing community PRs onto the project board (#23796)

2026-09-28 22:15:36 -07:00

hourly-mac-build.yml

fix(macos): preserve terminal permissions across app updates with a small signed helper (#27021)

2026-10-11 14:54:36 -04:00

issue-os-labeler.yaml

perf(ci): reduce queue pressure without paid runners (#23053)

2026-09-25 22:48:28 -07:00

linux-wayland-gpu-sandbox.yml

build: upgrade to pnpm 12 (#17156)

2026-08-29 14:13:26 -07:00

macos-updater-tests.yml

ci: skip checks for draft pull requests [skip ci] (#29484)

2026-10-15 23:18:13 -07:00

mobile-android-release.yml

refactor(mobile): every build is the OTA shell; drop the native/OTA build switch (#29969)

2026-10-17 02:12:06 -07:00

mobile-ios-release.yml

Pin CocoaPods release installer in the Ruby lockfile (#30181)

2026-10-17 22:00:57 -07:00

mobile.yml

fix(mobile): break the transport and terminal-document import cycles, and gate cycles after mobile's install (#30290)

2026-10-18 04:00:24 -07:00

node-next-compat.yml

perf(ci): plan the unit shards before the static-analysis gate instead of behind it (#23743)

2026-09-28 19:22:10 -07:00

node-server-tests.yml

feat(orcad): ship the mobile web page so phones paired to orcad get it (#29882)

2026-10-16 19:29:41 -07:00

packaged-browser-e2e.yml

refactor(build): relay survivors named for what they are (#27962)

2026-10-12 04:09:18 -07:00

performance-contracts.yml

ci: skip checks for draft pull requests [skip ci] (#29484)

2026-10-15 23:18:13 -07:00

pi-owner-runtime.yml

ci: skip checks for draft pull requests [skip ci] (#29484)

2026-10-15 23:18:13 -07:00

pi-provider-runtime.yml

ci: skip checks for draft pull requests [skip ci] (#29484)

2026-10-15 23:18:13 -07:00

pr-test-loc.yml

Reduce redundant CI runs, pnpm uploads, and fixture startups (#23145)

2026-09-26 01:17:51 -07:00

pr.yml

refactor(managed-server): move the managed server's install and deploy code into src/main/managed-server/install/ (#30335)

2026-10-18 05:20:19 -07:00

preflight-benchmark.yml

ci: skip checks for draft pull requests [skip ci] (#29484)

2026-10-15 23:18:13 -07:00

preflight-cli-platforms.yml

ci: skip checks for draft pull requests [skip ci] (#29484)

2026-10-15 23:18:13 -07:00

preflight-critical-path-benchmark.yml

ci: skip checks for draft pull requests [skip ci] (#29484)

2026-10-15 23:18:13 -07:00

pullfrog.yml

Pause Pullfrog reviews while the CI runner queue recovers (#25760)

2026-10-05 20:49:20 -07:00

readme-downloads-badge.yml

Use generated README downloads badge

2026-06-16 11:09:40 -07:00

release-cut.yml

fix(release): publish tags when main requires pull requests (#29985)

2026-10-16 22:57:25 -07:00

release-javascript-benchmark.yml

ci: skip checks for draft pull requests [skip ci] (#29484)

2026-10-15 23:18:13 -07:00

release-javascript.yml

perf(ci): compile release JavaScript once for all packaging hosts (#25828)

2026-10-06 13:22:04 -07:00

release-mac-build.yml

fix(macos): preserve terminal permissions across app updates with a small signed helper (#27021)

2026-10-11 14:54:36 -04:00

release-policy.yml

feat: live updates for agent state rules (#24387)

2026-10-02 14:59:24 -04:00

release-ref-validation.yml

ci: skip checks for draft pull requests [skip ci] (#29484)

2026-10-15 23:18:13 -07:00

skill-update-roundtrip.yml

ci: skip checks for draft pull requests [skip ci] (#29484)

2026-10-15 23:18:13 -07:00

ssh-hostile-hosts.yml

refactor(ssh): move the ssh-connection and ssh-target cluster into ssh/connection/ and ssh/targets/ (#30479)

2026-10-18 19:23:21 -07:00

ssh-windows-hosts.yml

refactor(ssh): move the ssh-connection and ssh-target cluster into ssh/connection/ and ssh/targets/ (#30479)

2026-10-18 19:23:21 -07:00

terminal-ime-e2e.yml

refactor(build): relay survivors named for what they are (#27962)

2026-10-12 04:09:18 -07:00

terminal-perf.yml

build: pin pnpm 12.14.0 (#30481)

2026-10-18 19:27:33 -07:00

unit-module-cache-benchmark.yml

Reduce CI import, frame-capture, teardown, and shard costs (#28879)

2026-10-14 17:10:55 -07:00

unit-plan.yml

perf(ci): plan the unit shards before the static-analysis gate instead of behind it (#23743)

2026-09-28 19:22:10 -07:00

unit-selection-evidence.yml

ci: reuse bundled Node for selection evidence (#29017)

2026-10-14 23:50:49 -07:00

unit-tests.yml

ci: move PR analysis and Linux unit tests to Blacksmith (#29362)

2026-10-15 21:07:49 -07:00

win-crash-survival-e2e.yml

ci(windows): build the agent hook before packaging in installer e2e workflows (#28137)

2026-10-12 19:04:02 -07:00

win-orcad-serve-switch-e2e.yml

feat(orcad): ship the mobile web page so phones paired to orcad get it (#29882)

2026-10-16 19:29:41 -07:00

win-update-e2e.yml

build: upgrade to pnpm 12 (#17156)

2026-08-29 14:13:26 -07:00

win-update-survival-e2e.yml

ci(windows): build the agent hook before packaging in installer e2e workflows (#28137)

2026-10-12 19:04:02 -07:00

windows-native-preparation-benchmark.yml

ci: reuse prepared Windows security addon in PR packages (#29352)

2026-10-15 21:05:34 -07:00

windows-signing-rehearsal.yml

perf(claude): run the Windows Claude hook as a signed native claude-hook.exe (#27719)

2026-10-12 16:09:30 -07:00

windows-terminal-restart-e2e.yml

fix(windows): preserve relocated terminals and native process scans (#22872)

2026-09-26 14:31:07 -07:00

windows-typecheck-parity.yml

ci: skip checks for draft pull requests [skip ci] (#29484)

2026-10-15 23:18:13 -07:00

windows-wsl-e2e.yml

Enable Claude plugin images and bounded terminal file transport (#28200)

2026-10-15 21:43:48 -07:00

wsl-orcad-e2e.yml

refactor(managed-server): move the managed server's install and deploy code into src/main/managed-server/install/ (#30335)

2026-10-18 05:20:19 -07:00

zcode-account-contract.yml

ci: reduce runner occupancy in release and contract jobs (#28930)

2026-10-14 19:53:57 -07:00

zcode-credential-contract.yml

ci: skip checks for draft pull requests [skip ci] (#29484)

2026-10-15 23:18:13 -07:00

zcode-transcript-contract.yml

ci: skip checks for draft pull requests [skip ci] (#29484)

2026-10-15 23:18:13 -07:00