Review HCAvailable condition for holistic expectation day 2

Most of our conditions status is driven by programatic output of reconciliation loops.

E.g: the HostedCluster available

• depends on kas, etcd and infra conditions.
• For kas/etcd we check the Deployment/stateful resource healthy

This is a good signal for day 1, but we might be missing relevant real state of the world for day 2. E.g:

• Do we flip HCAvailable condition if the our ingress controller is deleted/unhealthy.
• Do we flip HCAvailable condition if a Route resource is deleted?
• Do we flip HCAvailable condition if the LB is deleted out of band?

DoD:

Reproduce and review behaviour the examples above.

Consider adding additional knowledge for computing the HCAvailable condition. Health check on expected day 2 holistic e2e behaviour rather than in particular status of subcomponents.

E.g. actually query the kas through the url we expose