Try to reconcile license even in absence of known health status

[pebrc]

Fixes #6274

Change the logic so that license reconciliation is attempted as soon as we have endpoints in the K8s service. But only log an error if we think based on the observed cluster health that the license API call should succeed.

• in case of an expired license the observed health will be 403, and we will try to reconcile the license
• attempts will be logged by the es client in debug mode
• if the cluster should respond to requests based on the health status additional logging on errors will happen and a Kubernetes event will be generated. This restriction is motivated by the fact the users get confused by messages like "could not reconcile license" when in fact the underlying issue is that their cluster is down. So this preserves the functionality that Stricter notion of esReacheable: require health response #5796 introduced.

[thbkrkr]

LGTM

I kind of tested it by tweaking the operator (observer health / license reconciliation error).

[pebrc]

💚 All backports created successfully

Status	Branch	Result
✅	2.6

Questions ?

Please refer to the Backport tool documentation