Documentation
Reliability

Backups, restore, and PITR

Protect data and prove it can be recovered.

8 minute read ยท Updated September 2026

Protection layers

  • Scheduled encrypted backups provide durable recovery points.
  • Continuous WAL archiving supports point-in-time recovery within the retention window.
  • Off-server storage protects against a single host failure.
  • Recovery drills prove that artifacts are readable and procedures work.

Restore safely

  1. 1Open Backup & restore and select the source database.
  2. 2Choose a restore point or a timestamp inside the PITR window.
  3. 3Restore to a new isolated database first.
  4. 4Validate schema, row counts, critical records, and application behavior.
  5. 5Plan a controlled cutover; do not overwrite the source during investigation.

Recovery evidence

Record the restore start time, completion time, selected recovery point, integrity checks, RPO achieved, RTO achieved, and reviewer. Use Reliability for drill history and Launch readiness for unresolved controls.

Backups are not evidence by themselves

A successful upload proves storage, not recovery. Schedule recurring restore tests.

Was this guide useful?

If anything is unclear or incorrect, tell the support team.

Get help