Add liveness checks to clustermesh deployment

[nathanperkins]

Use the existing HTTP health check endpoints to ensure liveness of etcd and apiserver containers.

• etcd - https://etcd.io/docs/v3.4/op-guide/monitoring/#health-check
• clustermesh-apiserver - https://github.com/cilium/cilium/blob/080857bdedca67d58ec39f8f96c5f38b22f6dc0b/clustermesh-apiserver/main.go#L257

Tested locally on a kind cluster:

1. Create kind cluster via cilium instructions
2. Enable cilium with ./cilium --cluster-name=kind --cluster-id=0.
3. Enable clustermesh with ./cilium clustermesh enable --service-type ClusterIP
4. Verify clustermesh is available via ./cilium status output.

nathanperkins@nathanperkins1:~/git/cilium-cli (liveness)$ ./cilium status
...
Deployment        clustermesh-apiserver    Desired: 1, Ready: 1/1, Available: 1/1
...

[maintainer-s-little-helper]

Commit 171c522 does not contain "Signed-off-by".

Please follow instructions provided in https://docs.cilium.io/en/stable/contributing/development/contributing_guide/#developer-s-certificate-of-origin

[nathanperkins]

Actually the apiserver started to crashloop from failed queries to etcd. It is probably just the kind setup I'm using for testing, though. I will try again and if that doesn't work, I will try a different set up.

[nathanperkins]

I see that the etcd liveness check is failing because it should be HTTPS, but when changed to HTTPS I get this error:

Warning Unhealthy 58s (x3 over 78s) kubelet Liveness probe failed: Get "https://10.244.2.42:2379/health": remote error: tls: bad certificate