-
Notifications
You must be signed in to change notification settings - Fork 4k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
VPA admission controller can't admit pods in kube-system namespace #7392
Comments
/area vertical-pod-autoscaler |
Can you paste the output of the webhook: Do other webhooks work in AKS? (ie: is this AKS blocking changes to kube-system?) |
Here is the webhook
I check the other webhook in AKS but they seems not related to it |
That isn't a webhook for the VPA.
But the one you listed isn't it. |
Sorry I pasted the wrong one:
|
It seems like AKS adds some specific AKS config there:
Can you show the details of the namespace? |
It seems to need add a label in MutatingWebhookConfiguration in AKS. But when I edited |
I'm not sure what label you're talking about. I assume AKS is specifically designed to not allow VPA to modify kube-system resources. |
Here is the output:
|
It seems like the VPA is being configured to avoid mutating pods in the kube-system namespace. |
/remove-kind bug |
Yes, they have aks managed vpa and it need to add a label to webhook. Do upstream VPA can edit MutatingWebhookConfiguration to add a label? It will recreate a default one when recreate admission-controller pod |
like this:
|
Which variant of the VPA are you running? One installed using AKS or did you install this yourself? |
Ah, I see you're referring to https://learn.microsoft.com/en-us/azure/aks/faq#can-admission-controller-webhooks-impact-kube-system-and-internal-aks-namespaces- I guess you can make a PR to include a flag to toggle that |
Another option is to disable the webhook being created |
Yes, the config with that label is AKS VPA. I will try |
Hi, @adrianmoisey I have make a pr for it: #7402. Could you please have a look? Thanks! |
/close as #7402 has been merged. Thanks! |
/close |
@voelzmo: Closing this issue. In response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. |
Ref: - AKS: kubernetes#7392 - EKS: kubernetes#6915
Which component are you using?:
vertical-pod-autoscaler
What version of the component are you using?:
Component version: v1.2.1
What k8s version are you using (
kubectl version
)?:kubectl version
OutputWhat environment is this in?:
AKS
What did you expect to happen?:
The pod in kube-system namespace can change memory and cpu request
What happened instead?:
They just restart all time. Admission controller can't admit pod in kube-system namespace. But recommander and update can listen them. So they evict pod all the time but no changes.
How to reproduce it (as minimally and precisely as possible):
Anything else we need to know?:
Admission controller logs:
Recommender logs:
Updater logs:
The text was updated successfully, but these errors were encountered: