Question about "monitor cluster component"


I downloaded the metric server using the link in the lab, and I encountered the following error when I run “kubectl top nodes”

Sometimes, I can get the resource usage status, but I got this error sometimes

How can I solve this?

Thank you.

Hello, @hlngai
It takes time to collect metrics from the cluster. Is it after a few min ?


For the lab, I can get the resource usage after several min.

But for my testing environment, sometime I can get the resource usage, but sometimes I get this error

Might be a authentication or permission issue is the reason of fails. What about logs of metrics pods?

That’s quite weird, I try to get the log by the following command:
kubectl logs pod/metrics-server-xxxxx -n kube-system

The error is the following:
Error from server: Get https://xxx.xx.xx.xx:10250/containerLogs/kube-system/metrics-server-78457c557c-sp7km/metrics-server: dial tcp xx.xx.xx.xx:10250: i/o timeout