Handle cluster Environments not loading better on the FE
Motivation
This is a follow-up that came from: https://gitlab.com/gitlab-com/gl-infra/infrastructure/-/issues/9440.
When the cluster has a misconfiguration, for instance if a certificate got rotated and GitLab cannot anymore talk to it, the "Operation->Environment" page will show environments loading forever and will never show any error message indicating what's happened there. We found out this issue by finding the logs on Kibana.
Another problem is when the environments ReactiveCache
limit goes over its accepted limit. Then the dashboards will load forever without a clear reason to the user.
How to reproduce
- Associate a cluster and do some deployments to it so you can see them on the Deploy Boards.
- Edit your cluster details and update the certificate to something invalid.
- Comeback to the Deploy Boards and you won't be able to load the environments anymore.
Proposal
- Check cluster connectivity.
- Notify the FE that there are connectivity problems so we show an error messaged to the user.
- Notify the FE if
ReactiveCache
limits got reached
We already check for cluster connectivity on the cluster's detail page. Same logic should be applied for the Deploy Boards.