CloudNativePG Backup and Restore
CloudNativePG, together with the Barman Cloud Plugin (cnpg-system), continuously protects every managed PostgreSQL cluster (e.g. Keycloak’s database) against data loss. This scenario implements [RE02 Recoverability / Backup] .
Figure 1. Continuous WAL archiving, a scheduled base backup, and a point-in-time restore.
- Motivation
-
Because WAL archiving is continuous, the recovery point objective is minutes, not the interval between base backups. Restores create a new
Clusterresource from object storage rather than mutating a running one, so a botched restore never endangers the still-running original cluster.
See Data Protection for the underlying concept, including which clusters are covered and retention settings.