planning-disaster-recovery

Configure pgBackRest with continuous WAL archiving to S3, GCS, or Azure, then schedule weekly full and daily differential backups. Restore to a specific moment using pgbackrest with the recovery target time option for PITR.

Velero is the primary tool for Kubernetes DR, supporting persistent volume snapshots, namespace-scoped backups, and scheduled jobs. For control plane recovery, combine it with periodic etcd snapshots using etcdctl.

RTO is the maximum acceptable downtime before service must be restored, while RPO is the maximum acceptable data loss measured in time. Mission-critical Tier 0 systems typically require RTO under 1 hour and RPO under 5 minutes.

Yes, Velero supports restoring backups into clusters in different regions when the backup storage location is accessible. It also supports namespace remapping and storage class remapping during selective restores.

Run chaos engineering tests in staging using tools like Chaos Mesh or the included failover test scripts that stop the primary database and measure promotion time. Schedule monthly DR drills and restore backups to isolated staging environments for validation.

HIPAA requires 6 years of retention with encryption, while GDPR typically requires 1-7 years with EU data residency and right-to-erasure support. Implement retention with S3 or GCS lifecycle policies and immutable storage for ransomware protection.