Repository navigation
Incident 2026-10-20: production health check failed #99
Description
Incident: production health check failed
The scheduled health check for https://wulfland.github.io/Repstack/ failed.
Nobody wrote this task by hand; health-check.yml opened it from the
template in .github/INCIDENT_TASK.md.
| Detected | 2026-10-20 (UTC) |
| Failing health-check run | https://github.com/wulfland/Repstack/actions/runs/39395289877 |
| Latest deploy run | https://github.com/wulfland/Repstack/actions/runs/39395120852 |
| Deployed commit | 27feb69 |
| Last green deploy run | https://github.com/wulfland/Repstack/actions/runs/39268902273 |
| Last green commit | 7745b60 |
| Automatic rollback | Started a redeploy of the last green commit 7745b606b7ba2502da48bba973e31c2063ce83c9 (deploy run: https://github.com/wulfland/Repstack/actions/workflows/deploy.yml?query=event%3Aworkflow_dispatch). The next health checks show whether that restored service. |
Task
Work strictly within the Incident policy in AGENTS.md. The pipeline
already took care of the rollback (see the table). Do not redeploy or
dispatch workflows yourself.
- Diagnose. Read the logs of the failing health-check run and the latest
deploy run above. Find the deploy or change that broke production and say
why it broke it. - Check the rollback. The comments below list every health check that
failed after this issue was opened. If the rollback ran and later checks
stop failing, service is restored. If it was skipped, failed, or
production stays red, say so in the PR so a human can step in. If a
feature flag caused the break, turn it off (defaultfalse) in your fix.
Do not touch the Dexie schema, secrets or repository settings. - Fix permanently. Open a pull request with the root-cause fix and an
incident summary atincidents/2026-10-20.md(template in
incidents/README.md), linking the failing runs above and the rollback.
Reference this issue in the PR.
Merging the fix PR closes this issue. Before it is merged, the health check
must be green again.
Activity