Reliability

Recovery belongs in the control plane, not a private runbook.

Backup creation, safety snapshots, restores, point-in-time recovery and deployment history are explicit operations with visible outcomes.

verified backupspre-restore snapshotsPITR workflows

Operational history

When state changes, the event remains visible.

Deployments and recovery actions share an activity history, so operators can trace what changed and when without relying on shell history.

Create and verifyBackups are treated as recoverable artifacts, not only scheduled jobs.

Snapshot before restoreCapture the state being replaced before a destructive recovery step.

Restore with contextRecovery events remain alongside deployment and operator activity.

instance / activity

postgres / data

Data service

Database inspection, backups, networking, metrics and service settings stay attached to the datastore instead of being spread across unrelated tools.

Create & verifyintegrity is part of the workflow

verified

Capture safety snapshotpreserve the state being replaced

created

Restorerecord the recovery event

restored

Backup destinations, jobs and restore operations are platform concepts instead of ad-hoc scripts.

Recovery workflows preserve context around when and why a datastore is moved backward in time.

Confirmation steps and preconditions reduce the chance that recovery work creates a second incident.

Prepare before the incident

Use the backup and PITR guides to set destinations, schedules and restore procedures.