Staging Deployment

Rollback Checklist

Rollback is a human-approved operational decision. Sprint 23 adds no destructive rollback automation.

Use this when...

Use this when you need step-by-step confirmation before or after an operation.

Rollback is a human-approved operational decision. Sprint 23 adds no destructive rollback automation.

Roll Back When

  • Login or core dashboard routes are unavailable.
  • Readiness fails after deployment.
  • A migration causes unexpected critical data behavior.
  • A safety boundary is accidentally enabled.

Code Rollback

  1. Identify the last known good commit/image.
  2. Get approval from the accountable owner.
  3. Redeploy the previous app image or branch.
  4. Verify /api/health, /api/ready, login, and smoke tests.

Database Rollback

Prefer fix-forward when safe. Restore from backup only after approval and after confirming the target database. Document the backup used, restore time, approver, and validation results.

Feature Flag Safety

Disable risky flags first when possible:

  • external sending/sync/payment flags
  • autonomous action flags
  • future integration placeholders

Document the rollback in System QA or the deployment notes.

Developer Details
Source
apps/docs/content/staging-deployment/rollback-checklist.md
Owner
Platform / Operations
Status
Current
Module
Staging Deployment
Related route
/api/health