Cluster Architecture:

1. Private autopilot cluster with no external endpoint
2. For connecting to api-server, using jump-host, which only is a way to connect to cluster and run kubectl commands.
   below is the attached details of cluster and service account.

service-account-details

Issue:

I have made a deployment of my application, but when I try to see logs or run “exec” command I get error saying :

[ Error from server: Get “https://10.250.82.36:10250/containerLogs/qa1/app1-5dd86f95dc-dmgjj/app1”:
dial tcp 10.250.82.36:10250: i/o timeout ]

I am unable to understand what the issue is.

Earlier when i was trying to make deployment of one more service, I was getting same error, but after doing some R&D, I was able to resolve the issue by creating a serviceAccount and associate that to service-deployment. With this I was able to see logs or run “exec” command. This works once every few times and not for each invocation. The below screenshot shows more details about the installation and execution.

I have tried same solution for my main application deployment, but I am only able to see error messages. Below screenshot has more details about the error.

Let me know if any suggestions to exec or get logs for the pod.