Probably contrary statement regarding usage of multiple metrics server pods #1643
Description
What happened: Kubernetes Metrics Server (KMS) documentation has two statements when it comes to usage of multiple metrics server pods - Can I run more than one instance of metrics-server? which states but only one instance will be actively serving metrics API. and High Availability which states it is recommended to add the --enable-aggregator-routing=true CLI flag to the kube-apiserver so that requests sent to Metrics Server are load balanced between the 2 instances. This is somehow confusing for me.
What you expected to happen: Because KMS uses active-active, not leader election, and kube-apiserver with --enable-aggregator-routing=true uses endpoints to round-robin I expect that all pods are used. Is this understanding correct?
For example kube-controller-manager HPA controller with multiple threads will utilise all metrics server pods which leads to faster work_queue processing.
Anything else we need to know?:
Environment: EKS with multiple KMS replicas and kube-apiserver --enable-aggregator-routing=true
-
Kubernetes distribution (GKE, EKS, Kubeadm, the hard way, etc.): any
-
Container Network Setup (flannel, calico, etc.):
-
Kubernetes version (use
kubectl version): -
Metrics Server manifest
spoiler for Metrics Server manifest:
- Kubelet config:
spoiler for Kubelet config:
- Metrics server logs:
spoiler for Metrics Server logs:
- Status of Metrics API:
spolier for Status of Metrics API:
kubectl describe apiservice v1beta1.metrics.k8s.io/kind bug