# Runbook: Database Backup And Restore Drill

Last updated: 2026-06-24

## Purpose

Prove that encrypted PostgreSQL backups can be restored and validated within the agreed RPO/RTO.

## Inputs

- Backup reference and timestamp.
- Target restore environment.
- Expected RPO/RTO.
- Current migration plan reference.
- Operator and approver.

## Drill Steps

1. Select an encrypted backup or PITR restore point.
2. Restore into an isolated staging/preprod database.
3. Confirm network access is restricted to the drill operators.
4. Run migration status checks and compare with the release migration plan.
5. Run backend tests that validate ledger invariants and restored schema behavior.
6. Run application smoke checks for login, account list, transfer list, cards, FX and admin dashboard reads.
7. Run audit-chain verification if restored data includes audit hashes.
8. Record RPO, RTO, validation summary and evidence link in `database_restore_drill_evidence`.
9. Obtain approval from the database/platform owner.

## Pass Criteria

- Restore completes within RTO.
- Backup point satisfies RPO.
- Schema, ledger and audit validation pass.
- No secrets, customer PII or production data are exposed outside approved drill access.
- Evidence is linked from the production readiness or release record.

## Failure Handling

Open a P0 platform risk issue if restore fails, validation fails, evidence is incomplete or access controls are not enforced.