Recommended setup:

1. Make sure the metrics-server is installed, for example:
   • helm repo add metrics-server https://kubernetes-sigs.github.io/metrics-server/
   • helm repo update
   • helm upgrade --install --set args={--kubelet-insecure-tls} metrics-server metrics-server/metrics-server --namespace kube-system
2. Run ./redeploy-pod.sh pod-with-issue.yaml.
3. In a separate shell, start watch kubectl top pod --namespace=kubelet-stats-test and let it run.
4. In a separate shell, start watch ./print-kubelet-stats-memory-usage-bytes.sh and let it run. This prints the value that kubelet stats report (that is, kubectl get --raw /api/v1/nodes/docker-desktop/proxy/stats/summary -> pod -> .memory.usageBytes formatted in MB). Note: If your node name is not docker-desktop, you might need to edit this file.
5. Optional: In a separate shell, start k9s, open the namespace kubelet-stats-test and lists its pods to observer the memory values it reports. (Might use kubectl top pod or something similar under the hood, not sure.)

Observations:

• The memory usage of the pod as reported by kubectl top and k9s will be at 0 or close to zero.
• The memory.usageBytes reported by kubelet stats (/api/v1/nodes/docker-desktop/proxy/stats/summary) will be around 100 MB.

All three values represent the memory usage of the pod, but the reported value from kubelet stats is different.