Issue 102 - Allow setting runtimeClassName

[nvvfedorov]

No changes in the values.yaml, except comments added for better documentation.

Test steps:

1. Install K3S

curl -sfL https://get.k3s.io | K3S_KUBECONFIG_MODE="644" sh -s -

2. View available run times:

kubectl get runtimeclasses

You should expect to see:

NAME                  HANDLER               AGE
crun                  crun                  4h
lunatic               lunatic               4h
nvidia-experimental   nvidia-experimental   4h
slight                slight                4h
spin                  spin                  4h
wasmedge              wasmedge              4h
wasmer                wasmer                4h
wasmtime              wasmtime              4h
wws                   wws                   4h
nvidia                nvidia                4h

3. Deploy the dcgm exporter

helm install --generate-name ./deployment/ --set runtimeClassName=nvidia --set serviceMonitor.enabled=false

4. Check pod configuration:

kubectl get pods deployment-1710791202-dcgm-exporter-jrgc7 -o yaml

You should see: runtimeClassName: nvidia and that the pod is up and running.