Every agent pod write the following message each second in an EKS cluster:

<probe> WARN: 2022/12/05 11:05:33.161545 No node name and cannot obtain local pods, reporting all (which may impact performance): Get http://127.0.0.1:10255/pods/: dial tcp 127.0.0.1:10255: connect: connection refused

This is because EKS disabled the local kubelet endpoint because it is deprecated:

• https://kubernetes.io/docs/reference/command-line-tools-reference/kubelet/
• Disable the kubelet readonly port kubernetes/kubeadm#732
• Secure Kubelet's componentconfig defaults while maintaining CLI compatibility kubernetes/kubernetes#59666

There were discussions about this in the past:

• Weave scope running in eks throws error messages weaveworks/scope#3669
• Stop talking to the kubelet unsecured port weaveworks/scope#3242

The data should not be fetched via the unsecured endpoint.