Validate Cluster Monitor works in a "ship of Theseus" situation (all nodes replaced)

Especially with the Operator it's fairly common that all initial nodes in the cluster are replaced, such that none of the nodes in the cluster are at their original hostname (e.g. after a rolling upgrade).

We should validate that the Cluster Monitor manages to keep track of nodes in such a situation - ideally automatically as part of tests, but a manual validation is a good start too. If it does turn out to break, file follow-up CMOSes to fix.