Improve kubernetes alerting cluster driver status

[alexandreLamarre]

In the current implementation, alerting cluster shows status as Pending until all pods are in the OK state.

This isn't ideal since the UI depends on this status to interact with the other alerting APIs

• We should differentiate between waiting for startup status
• Once startup has finished, we should show OK as long as there is one pod in a healthy state ready to receive traffic for example in case one pod meets resource limits and crashes.