1

I run a small Kubernetes cluster on the managed Google Kubernetes Engine (GKE) service. I didn't to any further changes, except from deploying three simple deployments. So I assumed that everything, including metrics, works out of the box. Unfortunately, when I filter for metric-server, I see a lot of errors in stackdriver that do not seem right.

Does somebody know where these are coming from, what impact they have and how to fix it (although I would assume Google needs to fix I, since I pay for GKE :D ).

Thanks in advance.

GKE Kubernets GCP logs

Simon Buchholz
  • 11
  • 2
  • Hello. To be able to fully understand what is happening in your setup, please include information like: cluster version, what exactly have you deployed on the cluster. Also have you waited a bit for logs to be retrieved? Please have a look on this thread: https://github.com/kubernetes-sigs/metrics-server/issues/349#issue-518111196 – Dawid Kruk Jul 22 '20 at 07:36

1 Answers1

1

Edit the metrics-server-deployment and add:

command:
    - /metrics-server 
    - --kubelet-preferred-address-types=InternalIP
    - --kubelet-insecure-tls

Note that this not a very good solution from security standpoint.

Arghya Sadhu
  • 41,002
  • 9
  • 78
  • 107
  • 1
    I don't have a GKE cluster running at the moment, so can't confirm this: Does the metrics-server deployment have the reconcile: true annotation? If so, this answer won't work – Patrick W Jul 21 '20 at 13:35