What happened?

It seems the apiserver will fail bootstrap if the etcd override endpoint is not healthy.
But after the bootstrap completes, if the etcd override endpoint become unhealthy, the apiserver health check will still report OK while kubectl get cs will report etcd override endpoint is not healthy.

What did you expect to happen?

APIserver health check should report unhealthy when an etcd override endpoint is unhealthy.

How can we reproduce it (as minimally and precisely as possible)?

Run 2 etcd clusters: one for events and one for the other resources. Then configure apiserver to use the event etcd using the --etcd-servers-overrides flag. After the apiserver complete the bootstrap, kill the event etcd.
You will be able to see APIServer health check still reporting OK.

Anything else we need to know?

No response

Kubernetes version

1.32. I think I have seen this issue in older version as well.

Cloud provider

N/A

OS version

# On Linux:
$ cat /etc/os-release
# paste output here
$ uname -a
# paste output here

# On Windows:
C:\> wmic os get Caption, Version, BuildNumber, OSArchitecture
# paste output here

Install tools

Container runtime (CRI) and version (if applicable)

Related plugins (CNI, CSI, ...) and versions (if applicable)