How do I plan backups and disaster recovery (RPO and RTO)?
- RPO — the maximum tolerable data loss; RTO — the maximum tolerable time to recover. Set both per datastore by business criticality
- Automate encrypted backups and keep a copy in an independent failure domain
- Cover everything: the database, object storage, and secrets/configuration — not just SQL dumps
- Test the restore in an isolated environment and confirm it produces a *working application*; a successful backup is not proof of a successful restore
- Use point-in-time recovery where supported
- Plan the human side: break-glass access, incident communication and a migration rollback or roll-forward procedure
- Rehearse annually or more often — and after major infrastructure changes