Thank you for this helpful example.

I have Zalando postgres-operator running postgres HA on three microk8s nodes using microk8s-hostpath local storage. Zalando manages the postgres replicas, and failover works.

If I replace microk8s-hostpath local storage with rook ceph block storage, will that result in redundant replication? Meaning, will Zalando still maintain postgres replicas across the cluster on its own, even though rook ceph is now maintaining block level replicas?

Or do I need to somehow configure Zalando to rely on the underyling rook ceph HA, and no longer maintain its own postgres replicas?