POC Health checks that cause unpublishing of unhealthy endpoints

The first step here will likely be a spike / POC to see how this might function.

Early thoughts:
- Continue with the delete label and the stop labels as planned before
- Await visibility of label in zone before unpublishing (to avoid NXDomain)
- Process the unpublishing for healthchecks in a separate plan similar to the DNS Failover
- Only process unpublishing for healthchecks if no other changes occurred in this reconcile.
- HealthCheckProbe CRs created by kuadrant-operator
- Status of relevant DNSRecord updated from reconcile of HealthCheckProbeCR
- Some mechanism to avoid republishing unhealthy endpoints after unpublishing.

Start by reviewing the HealthCheckProbe CRs created by kuadrant-operator, the DNSRecord reconciliation that updates status, and the existing DNS Failover plan. Define and validate a POC covering label visibility, separate health-check unpublishing, reconcile safeguards, and prevention of republishing unhealthy endpoints.

Written by the indexing model from the issue text.