Add some explanation of the difference between the livez and readyz endpoints of the kubernetes API

[Ji-Elle]

What would you like to be added
The current documentation page regarding the Kubernetes API health endpoint is not clear about where the /readyz and /livez differ and in which situations one or the other should be called. It was already a point raised before and the conversation around the commit adding the current version of the documentation indicate that those endpoints may not be relied upon: if it is the case it could be a good thing to add a note in the documentation.

Why is this needed
When setting up a kubernetes cluster on premise you want to have multiple nodes hosting the control plane to help with high availability. Then when setting up some load balancers on front having them able to check which nodes are down so people would expect some easy healthcheck.

Comments
More generally the documentation is not clear (or lack locality of information) regarding what happens when you scale the api server: if I have 3 nodes hosting the API, is the information those health endpoints return only for the node receiving the call or for every node? What happens when 1 of those 3 nodes is down?

[k8s-ci-robot]

This issue is currently awaiting triage.

SIG Docs takes a lead on issue triage for this website, but any Kubernetes member can accept issues by applying the triage/accepted label.

The triage/accepted label can be added by org members by writing /triage accepted in a comment.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.