error: kubect top node

[auqf]

the metrics-server:0.8.0 can not support the kubectl top node? the metric-server log is normal, not error

[k8s-ci-robot]

This issue is currently awaiting triage.

If metrics-server contributors determine this is a relevant issue, they will accept it by applying the triage/accepted label and provide further guidance.

The triage/accepted label can be added by org members by writing /triage accepted in a comment.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.

[kincoy]

You can try running:
kubectl api-resources | grep NodeMetrics
If NodeMetrics is not listed, kubectl top node won’t work — even if metrics-server logs look fine. Also check if metrics.k8s.io shows up in kubectl get apiservices.

[dgsamper]

Same issue here. We updated our EKS to 1.33 and therefore the metrics server:

[dgsamper]

We were using the official EKS addon. That was our problem. Somehow, the eksbuild.2 image is bugged. We installed via helm chart and now it is working fine.

[auqf]

use the yaml to install on the k8s v1.33.0,it's still cannot get the node metrics, and cannot found any error logs

[kincoy]

use the yaml to install on the k8s v1.33.0,it's still cannot get the node metrics, and cannot found any error logs

Could you try running kubectl get --raw /apis/metrics.k8s.io/v1beta1/nodes | jq and kubectl get apiservices | grep "v1beta1.metrics.k8s.io"?

From what I see, the NodeMetrics API seems to be properly registered, I suspect there might be an issue with the metrics-server’s communication with the kubelet—possibly failing to scrape node metrics over the secure Kubelet endpoint (port 10250).
You might also try curling the kubelet’s /metrics/resource endpoint from within the metrics-server Pod to verify if it’s reachable.Additionally, consider increasing the metrics-server log verbosity by adding --v=6 or higher to its command line. This could provide more insight into why node metrics are not being collected.

[kincoy]

We were using the official EKS addon. That was our problem. Somehow, the eksbuild.2 image is bugged. We installed via helm chart and now it is working fine.

Looks like the FailedDiscoveryCheck was the root cause — it means the APIService couldn’t be properly discovered, so kubectl top node didn’t work.