I am struggling to debug an environment where cross-cluster communication seems to not be working (requests from the Rancher Dashboard aren’t working). I have noticed that the Cluster Agent pods status are in Crash Backoff. I am trying to understand the role of the cluster agent on a managed cluster that causes the Dashboard requests to a managed cluster to fail.