I'm currently running a Kubernetes cluster on Vagrant and currently testing the HPA using the metrics-server but my setup can't seem to read the metrics at all.
My setup consists of a single loadbalancer(HAProxy), a single Master Node, and two worker nodes.
So, I've pulled the log from the metrics-server pod using the following command:
kubectl logs -n kube-system metrics-server-85cc4d4cb9-sf8wh
and it shows the followings:
E0122 01:53:26.647172 1 sinkprov.go:135] duplicate pod kube-system/metrics-server-85cc4d4cb9-sf8wh received
E0122 01:54:26.653698 1 sinkprov.go:135] duplicate pod kube-system/metrics-server-85cc4d4cb9-sf8wh received
E0122 01:54:26.653838 1 sinkprov.go:135] duplicate pod kube-system/kube-proxy-dk9v8 received
E0122 01:54:26.653864 1 sinkprov.go:135] duplicate pod kube-system/weave-net-jkrn8 received
E0122 01:55:26.647113 1 sinkprov.go:135] duplicate pod kube-system/kube-proxy-dk9v8 received
E0122 01:55:26.647129 1 sinkprov.go:135] duplicate pod kube-system/weave-net-jkrn8 received
E0122 01:55:26.647133 1 sinkprov.go:135] duplicate pod kube-system/metrics-server-85cc4d4cb9-sf8wh received
E0122 01:56:26.646230 1 sinkprov.go:135] duplicate pod kube-system/weave-net-jkrn8 received
E0122 01:56:26.646245 1 sinkprov.go:135] duplicate pod kube-system/metrics-server-85cc4d4cb9-sf8wh received
E0122 01:56:26.646249 1 sinkprov.go:135] duplicate pod kube-system/kube-proxy-dk9v8 received
E0122 01:57:26.691044 1 sinkprov.go:135] duplicate pod kube-system/kube-proxy-dk9v8 received
E0122 01:57:26.691064 1 sinkprov.go:135] duplicate pod kube-system/metrics-server-85cc4d4cb9-sf8wh received
E0122 01:57:26.691069 1 sinkprov.go:135] duplicate pod kube-system/weave-net-jkrn8 received
I've tried to search for a similar posts on stackoverflow and github, but couldn't find any.
I hope you guys can help me figure this out.
p.s.
I've done the metrics-server argument setup as well adding --kubelet-insecure-tls
and more. I'm beginning to think that this is a Hypervisor network issue, or maybe not... I'm too much of a noob to know this at the moment :(
I'm starting to think that this was a CNI plugin problem. I've switched from using weavenet CNI plugin to a calico plugin and problem just went away :/... If any one of you can explain this symptom in a much detailed manner I'd really appreciate it! :)
Your errors seems to be similar to the one mentioned here- https://github.com/kubernetes-sigs/metrics-server/issues/179
Reason mentioned is - If you are using HPA, and if you have just set the resources in your HPA and you don't set the limits it will behave like this.