CloudNative PG

CloudNativePG is the Kubernetes operator that runs PostgreSQL on your own cluster. Edka installs it when you provision a PostgreSQL runtime through the Databases workspace and manages the cluster lifecycle through its CRDs.

• Runs 1, 3, or 5 instances with synchronous replication and automated failover
• Continuous backups to S3 compatible storage or Google Cloud Storage, with point in time recovery
• Restore to a new database from point in time recovery or a completed backup
• Connection pooling with a PgBouncer pooler
• Declarative roles, databases, extensions, pg_hba, pg_ident, and custom PostgreSQL parameters
• A data volume plus an optional dedicated pg_wal volume
• Runtime metrics: connection usage, throughput, cache hit ratio, and replica lag

How it fits in Edka

Edka installs cloudnative-pg as a PostgreSQL dependency alongside cert-manager, and adds barman-cloud-cnpg-plugin when backups are enabled. The catalog currently runs PostgreSQL 18.6, 18.3, 18.1, 17.11, 17.9, 16.15, 16.13, 15.19, and 15.17.

The operator comes from the upstream OCI chart at ghcr.io/cloudnative-pg/charts/cloudnative-pg. Edka installs it into the cnpg-system namespace, creates that namespace, and installs the CRDs. cert-manager goes in first because the operator uses it for its admission webhook certificates.

Verifying it works

The operator is ready once its pod is running and the CRDs are registered:

kubectl get pods -n cnpg-system
kubectl get crd | grep postgresql.cnpg.io

From there, a PostgreSQL runtime provisioned through the Databases workspace appears as a Cluster resource:

kubectl get clusters.postgresql.cnpg.io -A

Related docs

• Dedicated Databases Workspace
• Barman Cloud CNPG Plugin