Increased CPU usage from 21.07 to 21.10, no way to set limits

[mac-chaffee]

Describe the bug After upgrading Trident from 21.07.02 to 21.10.1, I'm experiencing higher CPU throttling in the CSI pods (the ones in the daemonset):

I have a default LimitRange which sets the CPU limit to 250m, but I can't seem to find out how to set the spec.containers.resources.limits.cpu amount when using the trident operator: https://netapp-trident.readthedocs.io/en/stable-v21.07/search.html?q=cpu

I don't want to increase the LimitRange since that would affect all containers in the namespace as well, which do not need the extra CPU.

Is it possible to control the CPU limit? If not, this would be a feature request.

Environment

• Trident version: 21.10.1
• Trident installation flags used: n/a
• Container runtime: containerd
• Kubernetes version: 21.1
• Kubernetes orchestrator: bare metal, kubespray
• Kubernetes enabled feature gates: n/a
• OS: RHEL8
• NetApp backend types: NAS

To Reproduce Upgrade from 21.07.02 to 21.10.1, observe CPU usage higher.

Expected behavior Similar CPU usage between versions, but if a new feature was added that did increase the CPU required, the default CPU limit/request would also be raised.

If there is no default CPU limit, users should be able to set one.

Additional context n/a

[mac-chaffee]

Looks like the daemonset definition comes from here, so I see it has no way to set the resource limits: https://github.com/NetApp/trident/blob/48af782eb8deb09256f710c876cc41ec37e0a757/cli/k8s_client/yaml_factory.go#L891

So this is officially a feature request. Apologies for mislabeling!