Implement cluster heatlh computation projection

[arbulu89]

Hey folks! This is the improvement for the cluster health computation. It is based on projections in the discovery side and the checks results event. Hopefully it makes the work.
I have written some E2E tests and all goes green (they are not uploaded yet).

Some more details:
The partial healths fields is a map which stores all the partial healths of a resource, cluster in our case (but easy extendable for other resources like sap systems, hosts, etc). So the partial health is updated in the different events (now, when a new cluster discovery is done and the cluster checks results are executed). The 2 healths are stored, and the overall health computed.

I have a doubt with the new table though (health_states). I thinking if we should truncate the values of this table on web service boot up. This would have 2 benefits:

• If for some reason we need the change the partial healths labels (hana_sr_health and config_checks by now), we would get rid off them, and we wouldn't have zombie entries.
• We wouldn't have "old" health states. Maybe it is safe to restore the health to nothing on boot up, to not show old values, as it might be misleading.

[fabriziosestito]

LGTM, I think it is the best trade-off with the current architecture.
Not sure about truncating the table at boot i think we could merge this and then have a discussione on it.

[nelsonkopliku]

LGTM as well.
Generally speaking I'd strive to find a clearer naming for things like HANASrHealth4 😄
but it is good enough here.

Not sure also about truncating health on boot, but let's discuss on that.

[arbulu89]

LGTM as well. Generally speaking I'd strive to find a clearer naming for things like HANASrHealth4 smile but it is good enough here.

Not sure also about truncating health on boot, but let's discuss on that.

Update at: 1718d47