search

prometheus-community / helm-charts

Prometheus community Helm charts
Apache License 2.0
5.12k stars 5.03k forks source link

[kube-prometheus-stack] alertmanager-kube-prometheus-stack-alertmanager-0 spawning in wrong node node #3592

Open kaykhan opened 1 year ago

kaykhan commented 1 year ago

Hi we use kube-prometheus-stack and we tried to modfy a node group recently and cluster-autoscaler reported, this is the first time seeing this error:

1 cluster.go:162] node ip-10-0-4-83.us-east-2.compute.internal cannot be removed: pod with local storage present: alertmanager-kube-prometheus-stack-alertmanager-0

We are confused why it was spawned on this node-group in the first place becuase we have the following set: Is there some other nodeSelector we must modify to get the above pod to spawn in the correct node group?

alertManager:
  alertManagerSpec:
    nodeSelector:
      acme/node-type: ops
stale[bot] commented 1 year ago

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Any further update will cause the issue/pull request to no longer be considered stale. Thank you for your contributions.