Operator updates services on each reconciliation

[pebrc]

While working on #5651 I noticed that we update the services created by the operator on every single reconciliation.

This seems to be caused by a diff in ClusterIPs where we ignore ClusterIPs when they are not a parseable IP as it is the case for None.

I don't recall why we have this special casing of non-IPs, I suspect a validation in an older version of Kubernetes might have been at play as I don't think it is there without reason. In any case we should revisit this and see if we can avoid unnecessary updates.

This needs to be tested across k8s versions probably to avoid any breakage.