Hi! Recently I've upgraded Keda to the 2.14.0 version using the latest Helm charts available. Everything deployed just fine but while looking at the logs of the metrics API server the following message kept appearing:
Failed to make webhook authorizer request: subjectaccessreviews.authorization.k8s.io is forbidden: User "system:serviceaccount:keda-operator:keda-operator" cannot create resource "subjectaccessreviews" in API group "authorization.k8s.io" at the cluster scope
While having a look at the keda operator clusterrole template in the main branch I've realized the resource is indeed missing from it. Not sure if related but after the upgrade all our ScaledObjects and Jobs can no longer scale properly which I assume (might be a wrong assumption) is because of this. By running kubectl describe on any of the HPAs associated with Keda resources we can see the following:
unable to get external metric ... unable to fetch metrics from external metrics API: an error on the server ("Internal Server Error: \"/apis/external.metrics.k8s.io/v1beta1/namespaces/...: subjectaccessreviews.authorization.k8s.io is forbidden: User \"system:serviceaccount:keda-operator:keda-operator\" cannot create resource \"subjectaccessreviews\" in API group \"authorization.k8s.io\" at the cluster scope") has prevented the request from succeeding (...)
Expected Behavior
The resource and necessary actions should be present in the clusterrole Helm template and message above shouldn't be present allowing for a normal scaling behavior using the latest version of the Helm chart.
Actual Behavior
Message keeps being printed by the metrics API server not allowing for the proper scaling of affected resources (ScaledObjects and ScaledJobs)
Steps to Reproduce the Problem
Deploy the latest version of the Helm chart
Check logs of API server
Describe HPAs associated with Keda scaling objects/jobs
Profit(?)
Specifications
KEDA Version:2.14.0
Platform & Version:linux/amd64
Kubernetes Version:1.28.9
Scaler(s):Please elaborate
PS: This seems to be very easy to fix and I have a PR ready, if bug is confirmed/reproducible by others I can take care of firing it.
Hi! Recently I've upgraded Keda to the 2.14.0 version using the latest Helm charts available. Everything deployed just fine but while looking at the logs of the metrics API server the following message kept appearing:
While having a look at the keda operator clusterrole template in the main branch I've realized the resource is indeed missing from it. Not sure if related but after the upgrade all our ScaledObjects and Jobs can no longer scale properly which I assume (might be a wrong assumption) is because of this. By running kubectl describe on any of the HPAs associated with Keda resources we can see the following:
Expected Behavior
The resource and necessary actions should be present in the clusterrole Helm template and message above shouldn't be present allowing for a normal scaling behavior using the latest version of the Helm chart.
Actual Behavior
Message keeps being printed by the metrics API server not allowing for the proper scaling of affected resources (ScaledObjects and ScaledJobs)
Steps to Reproduce the Problem
Specifications
PS: This seems to be very easy to fix and I have a PR ready, if bug is confirmed/reproducible by others I can take care of firing it.