Ignoring some fields' differences to avoid uneeded pods switchover

Please, answer some short questions which should help us to understand your problem / question better?

• Which image of the operator are you using? ghcr.io/zalando/spilo-15:3.0-p1
• Where do you run it - cloud or metal? Kubernetes or OpenShift? Bare Metal K8s
• Are you running Postgres Operator in production? Not yet
• Type of issue? Bug report, question

Every 10/30 minutes my DB pods are restarted which make the DB unavailable for a couple seconds. I think I've traced the issue
to the way the operator compares the pod in the cluster and what it expects to find.
I have a software running in my cluster that patches pods images to redirect it to a caching proxy. (say docker.io/library/toto -> mycachingproxy/docker.io/library/toto) which I think causes this.

So my question is : is my analysis correct ? I'd be glad to post a MR to fix this issue if that's the case.

Thanks by advance.

Logs from the operator :

postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:00:31Z" level=info msg="cluster has been synced" cluster-name=systems-services-harbor/harbor-db pkg=controller worker=0
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:29:53Z" level=info msg="SYNC event has been queued" cluster-name=systems-services-harbor/harbor-db pkg=controller worker=0
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:29:53Z" level=info msg="there are 1 clusters running" pkg=controller
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:29:53Z" level=info msg="syncing of the cluster started" cluster-name=systems-services-harbor/harbor-db pkg=controller worker=0
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:29:53Z" level=debug msg="team API is disabled" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:29:53Z" level=debug msg="team API is disabled" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:29:53Z" level=info msg="syncing secrets" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:29:53Z" level=debug msg="syncing master service" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:29:53Z" level=debug msg="final load balancer source ranges as seen in a service spec (not necessarily applied): [\"127.0.0.1/32\"]" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:29:53Z" level=debug msg="syncing replica service" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:29:54Z" level=debug msg="syncing volumes using \"pvc\" storage resize mode" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:29:54Z" level=info msg="volume claims do not require changes" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:29:54Z" level=debug msg="syncing statefulsets" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:29:54Z" level=debug msg="mark rolling update annotation for harbor-db-0: reason pod not yet restarted due to lazy update" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:29:54Z" level=debug msg="mark rolling update annotation for harbor-db-1: reason pod not yet restarted due to lazy update" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:29:55Z" level=debug msg="syncing Patroni config" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:29:55Z" level=debug msg="making GET http request: http://10.233.69.103:8008/config" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:29:55Z" level=debug msg="making GET http request: http://10.233.67.21:8008/config" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:29:55Z" level=debug msg="making GET http request: http://10.233.69.103:8008/patroni" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:29:55Z" level=debug msg="making GET http request: http://10.233.67.21:8008/patroni" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:29:55Z" level=debug msg="performing rolling update" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:29:55Z" level=info msg="there are 2 pods in the cluster to recreate" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:29:55Z" level=debug msg="subscribing to pod \"systems-services-harbor/harbor-db-0\"" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:30:11Z" level=info msg="pod \"systems-services-harbor/harbor-db-0\" has been recreated" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:30:11Z" level=debug msg="unsubscribing from pod \"systems-services-harbor/harbor-db-0\" events" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:30:11Z" level=debug msg="making GET http request: http://10.233.67.21:8008/cluster" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:30:11Z" level=debug msg="switching over from \"harbor-db-1\" to \"systems-services-harbor/harbor-db-0\"" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:30:11Z" level=debug msg="subscribing to pod \"systems-services-harbor/harbor-db-0\"" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:30:11Z" level=debug msg="making POST http request: http://10.233.67.21:8008/failover" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:30:13Z" level=debug msg="successfully switched over from \"harbor-db-1\" to \"systems-services-harbor/harbor-db-0\"" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:30:14Z" level=debug msg="unsubscribing from pod \"systems-services-harbor/harbor-db-0\" events" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:30:14Z" level=info msg="recreating old master pod \"systems-services-harbor/harbor-db-1\"" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:30:14Z" level=debug msg="subscribing to pod \"systems-services-harbor/harbor-db-1\"" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:30:30Z" level=info msg="pod \"systems-services-harbor/harbor-db-1\" has been recreated" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:30:30Z" level=debug msg="unsubscribing from pod \"systems-services-harbor/harbor-db-1\" events" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:30:30Z" level=debug msg="syncing pod disruption budgets" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:30:30Z" level=debug msg="syncing roles" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:30:30Z" level=debug msg="closing database connection" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:30:30Z" level=debug msg="syncing databases" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:30:30Z" level=debug msg="closing database connection" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:30:30Z" level=debug msg="syncing prepared databases with schemas" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:30:30Z" level=debug msg="syncing connection pooler (master, replica) from (true, nil) to (true, nil)" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:30:30Z" level=debug msg="final load balancer source ranges as seen in a service spec (not necessarily applied): [\"127.0.0.1/32\"]" cluster-name=systems-services-harbor/harbor-db pkg=cluster
postgres-operator-56fb8b4bdb-zp7nw postgres-operator time="2023-10-03T12:30:30Z" level=info msg="cluster has been synced" cluster-name=systems-services-harbor/harbor-db pkg=controller worker=0

Start by tracing the operator's “syncing statefulsets” and “performing rolling update” paths described in the logs, focusing on how pod differences trigger recreation. Confirm whether the caching-proxy image mutation is treated as drift, then add coverage showing that this mutation does not cause unnecessary pod recreation or switchover.