You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We configured HPA with object metrics type along with average value. We are noting scheduled scaler is failing to update HPA configuration and it would fail with following error:
E0117 00:30:00.053071 1 scaling-controller.go:171] FAILED TO UPDATE HPA: <HPA_NAME> - HorizontalPodAutoscaler.autoscaling "<HPA_NAME>" is invalid: spec.metrics[1].object.target.value: Invalid value: resource.Quantity{i:resource.int64Amount{value:0, scale:0}, d:resource.infDecAmount{Dec:(*inf.Dec)(nil)}, s:"0", Format:"DecimalSI"}: must be positive
Can we look into options to workaround the issue and add the necessary logic in scheduled-scaler code to handle this situation. Some work around has been mentioned in the issue kubernetes/kubernetes#87733
The text was updated successfully, but these errors were encountered:
After a quick review of the code I imagine the issue may be that it's using an older version of the api. I'm not sure if that's going to be a problem in general or just for that workaround you shared. I'd be curious to see what if that workaround sticks after getting updated by the scheduled scaler.
@ericuldall,
We are noticing this on the 1.29 version of the k8s cluster with HPA of v2 apiVersion.
Since I can patch the HPA object(adding a label) manually but schedulerScaler unable to make changes to minReplicas I thought of informing this here kubernetes/kubernetes#87733 (comment)
We configured HPA with object metrics type along with average value. We are noting scheduled scaler is failing to update HPA configuration and it would fail with following error:
It seems to be related to this K8s HPA bug kubernetes/kubernetes#87733
Until K8s bug is fixed, we will run into issues.
Can we look into options to workaround the issue and add the necessary logic in scheduled-scaler code to handle this situation. Some work around has been mentioned in the issue kubernetes/kubernetes#87733
The text was updated successfully, but these errors were encountered: