# Playbook: Accidental Delete / Data Corruption

**Triggers:** deleted repo, dropped table, bad migration, destructive script, bad Ansible run.

## Steps

1. **Stop writes** — scale deployment to zero, enable maintenance mode, or revoke DB write access.
2. **Assess scope** — what was deleted/corrupted, when, last known good state.
3. **Clone backup to staging** — never restore directly over production first.
4. **Identify last good point** — backup timestamp, git ref, DB snapshot.
5. **Restore selectively** — one repo, one table, or one PVC — not full blast restore unless required.
6. **Diff** — compare staging restore against current production; list deltas.
7. **Apply to production** — merge recovered data; re-run migrations if needed.
8. **Resume writes** — remove maintenance mode; monitor.
9. **Document** — `TestEvidence/` note with RPO achieved; update registers if backup gap contributed.

## Fast rollback

If local snapshot or recent git push exists, prefer that over full restore.

## References

`RTO-RPO-Matrix.md` · `BackupPolicy.md` · `AssetRegister.yaml`