CRD health check is not working reliably

[kahou82]

The current health check determine whether the CRD is degraded too early.

In this block of code, if the CRD is not ready and take some more time to register the endpoint in the k8s server, argocd sync will fail. This always happens if the cluster has a lot of CRDs to register.

If I restart the argocd sync, the degraded CRDs will become healthy.

Checklist:

• I've searched in the docs and FAQ for my answer: https://bit.ly/argocd-faq.
• I've included steps to reproduce the bug.
• I've pasted the output of argocd version.

Describe the bug

To Reproduce

Expected behavior

Screenshots

Version

Paste the output from `argocd version` here.

Logs

Paste any relevant application logs here.

[kahou82]

I will push a new PR to fix this issue