unable to fetch pod metrics

[RamanaChowdam]

i am installed metric server from aws documentation, when i look on metric server logs it giving some errors. but HPA is able to get resource utilization, and top pods also showing resources utilizaition.below are the error logs getting from metric server

E0929 10:10:07.183040 1 reststorage.go:160] unable to fetch pod metrics for pod default/podname: no metrics known for pod
E0929 10:10:07.187845 1 reststorage.go:160] unable to fetch pod metrics for pod default/podname: no metrics known for pod
E0929 10:10:07.189841 1 reststorage.go:160] unable to fetch pod metrics for pod default/pod: no metrics known for pod
E0929 10:10:37.653862 1 reststorage.go:160] unable to fetch pod metrics for pod default/pod: no metrics known for pod
E0929 10:10:37.655906 1 reststorage.go:160] unable to fetch pod metrics for pod default/pod: no metrics known for pod
E0929 10:10:52.669356 1 reststorage.go:160] unable to fetch pod metrics for pod default/pod: no metrics known for pod
E0929 10:10:52.671464 1 reststorage.go:160] unable to fetch pod metrics for pod default/pod: no metrics known for pod
E0929 10:11:08.083685 1 reststorage.go:160] unable to fetch pod metrics for pod default/pod: no metrics known for pod
E0929 10:11:08.086017 1 reststorage.go:160] unable to fetch pod metrics for pod default/pod: no metrics known for pod
W0929 13:05:44.203218 1 reflector.go:289] k8s.io/client-go/informers/factory.go:133: watch of *v1.Pod ended with: too old resource version: 4479009 (4521785)

[Cipher27]

+1 I am also facing this issue EKS 1.15, metrics server 0.3.1

Increasing verbosity to 10 still gives no more information than the line @RamanaChowdam put above. "no known metrics for pod"

Some pods are collected successfully. kubectl top pods and kubectl top nodes works fine.
As suggested in many of the other issues threads, any combination of the below added to the args of the metricserver deployment did NOT resolve the issue.

--kubelet-inescure-tls=true
--kubelet-preferred-address-types=InternalIP

What else can be an issue here? What else can I check?

[serathius]

"no metrics known for pod" doesn't imply any failures in Metrics Server. This log will be removed in next release #349

[osaffer]

Did you find solution ?

[serathius]

Those log lines were removed in new release as they don't indicate any problem. Please create separate issue and fill bug request form.