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
Describe the problem/challenge you have
[A description of the current limitation/problem/challenge that you are experiencing.]
Using the default velero helm chart resource requests results in OOMKilled for the velero container on my cluster.
I can override these resource requests, which is not an issue. The issue is that the defaults define CPU limits and have divergent memory requests and limits. I largely agree with the stance discussed https://home.robusta.dev/blog/stop-using-cpu-limits
I would like to set these to something akin to (although for my cluster slightly more memory):
resources:
requests:
cpu: 1000m
limits:
memory: 512Mi # this limit becomes a default request without repeating configuration and less error prone
I know of no way to set the resource requests and limits to the above, since I can't nullify the defaults of the helm chart itself. I am less worried about repeating memory limits in requests than I am about not being able to define no CPU limit.
Ideally, the helm chart would not come pre-loaded with its own resource requests, to also give it the best chance of running on minimal resource environments like Minikube.
Describe the problem/challenge you have
[A description of the current limitation/problem/challenge that you are experiencing.]
Using the default velero helm chart resource requests results in OOMKilled for the velero container on my cluster.

I can override these resource requests, which is not an issue. The issue is that the defaults define CPU limits and have divergent memory requests and limits. I largely agree with the stance discussed https://home.robusta.dev/blog/stop-using-cpu-limits

The current resource requests as are currently defined are:
https://github.com/vmware-tanzu/helm-charts/blob/bf3003e912af5c3828dfac4259154a71c58baab7/charts/velero/values.yaml#L76C1-L82C18
I would like to set these to something akin to (although for my cluster slightly more memory):
I know of no way to set the resource requests and limits to the above, since I can't nullify the defaults of the helm chart itself. I am less worried about repeating memory limits in requests than I am about not being able to define no CPU limit.
Ideally, the helm chart would not come pre-loaded with its own resource requests, to also give it the best chance of running on minimal resource environments like Minikube.
Bitnami has a more complex setup, but they enable overrides with some basic presents if that is still necessary without losing user overrides (taking one of their random charts):
https://github.com/bitnami/charts/blob/3117b573147401387321a2a0e1f587978cef0ed2/bitnami/ghost/values.yaml#L306C1-L317C14
Although, I do think the simplest solution is to not include defaults at all.
Thanks for reading!
Describe the solution you'd like
[A clear and concise description of what you want to happen.]
Default resource requests are completely overridable by having no defaults
Anything else you would like to add:
[Miscellaneous information that will assist in solving the issue.]
Environment:
Environment not relevant to this feature since it pertains to overridability.
helm version
):helm list -n <YOUR NAMESPACE>
):kubectl version
):/etc/os-release
):The text was updated successfully, but these errors were encountered: