Section 34 · API & Security

Rollback & Recovery

Deployment rollback, database migration rollback, and disaster recovery capabilities

8 items 5 critical 3 recommended

This guide walks you through auditing a project's rollback and disaster recovery capabilities - deployment rollback, database migration rollback, and full recovery from backups.

The Goal: Two-Minute Recovery

When things go wrong, speed matters. Every minute of downtime costs trust. This audit ensures you can undo mistakes fast.

  • Documented — Deployment rollback procedures can be executed quickly (under 2 minutes) by any team member
  • Migration-aware — Database rollback strategies exist, especially for destructive schema changes
  • Full-stack — Disaster recovery procedures cover restoring the entire system from backups
  • Tested — Rollback and recovery procedures have been validated and actually work
  • Objective-driven — Recovery objectives (RTO/RPO) are defined and achievable with current infrastructure

Before You Start

  1. Identify deployment platform (Vercel, Railway, Fly.io, K8s, custom CI/CD)
  2. Identify database and migration tool (Prisma, Drizzle, Knex, etc.)
  3. Identify backup strategy (provider snapshots, S3, PITR)
  4. Check for existing runbooks (disaster recovery documentation)