Incident Response

Severity

• SEV-1: broad outage or critical data/security risk
• SEV-2: major feature unavailable or heavily degraded
• SEV-3: partial impact with workaround

First 15 minutes

1. Confirm impact scope and affected environments.
2. Check recent deploys/config changes.
3. Validate health endpoints and dependencies.
4. Decide rollback vs runtime mitigation.

Mitigation options

• rollback latest deployment
• disable risky paths with runtime flags
• pause non-critical async processing
• switch provider path when single provider is degraded

Closeout

• verify core journeys are healthy
• publish postmortem with corrective actions
• update runbooks/alerts/tests

Detailed operations copy: docs/runbooks/incident-response.md