Support schedulerName and runtimeClassName in the operator Deployment

The operator chart's Deployment already exposes most pod scheduling knobs (`nodeSelector`, `topologySpreadConstraints`, `affinity`, `tolerations`, `priorityClassName`), but not `schedulerName` or `runtimeClassName`.

Both are standard PodSpec fields. `schedulerName` lets a cluster route the operator pod to a custom scheduler, and `runtimeClassName` is needed when the operator must run under a specific container runtime (for example gVisor or Kata on hardened nodes). Today neither can be set without post-render patching.

Proposed change, keeping current behavior by default:

- add `schedulerName: ""` and `runtimeClassName: ""` to `values.yaml`
- render them in `templates/deployment.yaml` next to the existing scheduling block, gated so an empty value emits nothing

Happy to open the PR if this sounds reasonable. Wanted to check on the approach and priority first, per the contributing guide.

Start with values.yaml and templates/deployment.yaml, locating the existing scheduling configuration in the Deployment template. Verify that schedulerName and runtimeClassName default to empty values, render only when configured, and preserve current behavior when unset; check the rendered chart output for both cases.

Written by the indexing model from the issue text.