This message was deleted.

The CPU Reserved metric is pointing to the the node's

status.requested

object. This would be the same as checking the "Allocated resources" when running

kubectl describe node <node-name>

- for example:

Allocated resources:
  (Total limits may be over 100 percent, i.e., overcommitted.)
  Resource           Requests    Limits
  --------           --------    ------
  cpu                700m (17%)  1 (25%)
  memory             240Mi (3%)  270Mi (3%)
  ephemeral-storage  0 (0%)      0 (0%)
  hugepages-2Mi      0 (0%)      0 (0%)

As for documentation, I wasn't able to find anything definitive on this, but from the kubectl source code, it appears to be all of the pod requests and limits of that particular node.

thank you so much @billions-vase-14972 ! So it has nothing to do with requests/limits set from Rancher on a project 🤔 ?

Correct, it wouldn't be based off of a project as this is a Rancher specific label and annotation

<http://field.cattle.io/projectId|field.cattle.io/projectId>

for Namespaces - As far as I can tell, it is specifically the requests/limits set on the containers of Pods.

much appreciated @billions-vase-14972 🙇 Clients set way too high requests while usage is next to nothing. I think we need implore them to set the requests lower and remove the limits where possible.

Glad to help. I'm not sure of your environment capabilities, but perhaps a LimitRange could be useful in this case.

thanks!!! Never worked with those, will do some reading 🙂