[RFC internal#219 Phase 4 PR-2] Flip continue-on-error true→false on stable CI jobs #623

Open
opened 2026-05-12 00:51:49 +00:00 by hongming · 1 comment
Owner

Follow-up to #622. RFC internal#219 Phase 4 step 2 — flip continue-on-error: true → false on the .gitea/workflows/ci.yml jobs that pass reliably on main. Approach: pull last 50 main-branch CI runs, compute per-job pass-rate, flip the ones >=98% green. Do NOT flip flaky jobs — file root-cause investigation issues for those instead (no fix-by-mute). Blocks: requires #622 to land first so all-required catches the flips. Memory ref: feedback_no_such_thing_as_flakes (every intermittent failure is a real bug). Test plan: after flip, open one PR that intentionally fails the flipped job, verify protection blocks merge; revert if any unexpected regressions on main.

Follow-up to #622. RFC internal#219 Phase 4 step 2 — flip continue-on-error: true → false on the .gitea/workflows/ci.yml jobs that pass reliably on main. Approach: pull last 50 main-branch CI runs, compute per-job pass-rate, flip the ones >=98% green. Do NOT flip flaky jobs — file root-cause investigation issues for those instead (no fix-by-mute). Blocks: requires #622 to land first so all-required catches the flips. Memory ref: feedback_no_such_thing_as_flakes (every intermittent failure is a real bug). Test plan: after flip, open one PR that intentionally fails the flipped job, verify protection blocks merge; revert if any unexpected regressions on main.
hongming added the tier:high label 2026-05-12 00:51:49 +00:00
core-devops was assigned by hongming 2026-05-12 00:51:49 +00:00
Member

Hourly CI/CD triage evidence (hongming-codex-laptop, 2026-05-12T19:50Z): lint-continue-on-error-tracking is now a current merge blocker on multiple PR heads.

Verified PR #755 run 18587: the lint emits untracked continue-on-error: true errors across workflows including e2e-staging-saas.yml, e2e-staging-sanity.yml, gate-check-v3.yml, handlers-postgres-integration.yml, harness-replays.yml, lint-workflow-yaml.yml, publish-canvas-image.yml, redeploy-tenants-on-main.yml, redeploy-tenants-on-staging.yml, staging-verify.yml, and others. The required fix class is tracker comments or removing the masks, not rerunning CI.

Hourly CI/CD triage evidence (hongming-codex-laptop, 2026-05-12T19:50Z): `lint-continue-on-error-tracking` is now a current merge blocker on multiple PR heads. Verified PR #755 run 18587: the lint emits untracked `continue-on-error: true` errors across workflows including `e2e-staging-saas.yml`, `e2e-staging-sanity.yml`, `gate-check-v3.yml`, `handlers-postgres-integration.yml`, `harness-replays.yml`, `lint-workflow-yaml.yml`, `publish-canvas-image.yml`, `redeploy-tenants-on-main.yml`, `redeploy-tenants-on-staging.yml`, `staging-verify.yml`, and others. The required fix class is tracker comments or removing the masks, not rerunning CI.
Sign in to join this conversation.
2 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: molecule-ai/molecule-core#623