Reliability center

Alert policies, incidents, and continuous proof that backups can actually be restored.

0Alert policies

0Open incidents

0Verified restores

PostgreSQL cluster

Primary health, WAL archiving, streaming replicas, and lag.

collecting0 streaming replicasHA activation pending standby deployment0 WAL segments archived

Encrypted off-site backup

Restic repository evidence from DigitalOcean Spaces.

collectingAwaiting next verified runNightly encrypted backup and repository integrity check

Recovery proof

Monthly full restore, PITR boundary, inventory, and replica verification.

collectingAwaiting evidenceOldest recovery boundary not yet publishedMonthly isolated proof

Independent HA witness

Quorum, fencing evidence, replay lag, and promotion authorization.

collectingAwaiting witness verification0 B replay lag · fence not active

Automatic failover gates

Every control must pass before unattended promotion can be enabled.

Full disaster-recovery exercise

Fresh backup → isolated restore → validation evidence → cleanup.

Database

Production traffic and credentials remain unchanged.

Run recovery drill

Creates a dump, restores it in isolation, validates it, then cleans up.

Source database

1. Backup2. Restore3. Validate4. Cleanup

Automated recovery drills

Continuously verify a fresh backup on an isolated database.

DatabaseFrequency

Scheduled validation

0 databases

Schedule validation for production databases.

Alert policies

Incidents

Everything is within policy.

Recovery drill history

Disaster-recovery evidence

0 exercises

Run one to capture RPO, RTO, restoration, and cleanup evidence.

HA failover history

0 events

The event ledger will record preflight, fencing, authorization, promotion, rejoin, and abort evidence.