CI Failure Response Runbook

Trigger

Use this runbook when a pull request or main branch pipeline fails.

Steps

1. Identify the failing job and capture the first actionable error.
2. Reproduce locally with the closest equivalent command.
3. Determine whether failure is code regression, flaky test, or environment issue.
4. Apply the smallest safe fix and rerun relevant checks.
5. If flaky, document evidence and open a follow-up task for stabilization.
6. Confirm all required checks pass before merge/release.

Escalation

• Escalate immediately if failure blocks releases or affects multiple repositories.
• Include failing job URL, error summary, and attempted mitigations in escalation.