K8SSAND-1806 ⁃ User keyspaces shouldn't be altered automatically during a DC decommission

[adejanovski]

Removing a DC from the K8ssandraCluster spec currently alters all keyspaces (system and user) to remove replicas from that DC, and then deletes the CassandraDatacenter object.
This can create bad surprises in multi tenant clusters where a team could delete a DC without asking all tenants if they're ok with it.
We want to add safety and prevent DC decommissions if any user keyspace still has replicas on it, informing the operators of the nature of the issue.
Since network calls are involved and due to the nature of multi cluster deployments, we cannot rely on webhooks for this and have to accept the patches. Instead, we'll create events to notify ops of the problem, and potentially enrich the K8ssandraCluster object status with the reconcile error message.

┆Issue is synchronized with this Jira Task by Unito
┆friendlyId: K8SSAND-1806
┆priority: Medium