Add per-state duration metric to reconciliation

[rajathagasthya]

Description

A reconcile processes many component states, but existing metrics only
expose whole-reconcile totals, so a slow state cannot be identified
from metrics alone. Add a gpu_operator_state_seconds histogram
labeled by controller and state, covering the ClusterPolicy,
GPUCluster, and NVIDIADriver paths.

Checklist

• No secrets, sensitive information, or unrelated changes
• Lint checks passing (make lint)
• Generated assets in-sync (make validate-generated-assets)
• Go mod artifacts in-sync (make validate-modules)
• Test cases are added for new code paths

Testing

go test ./controllers/ ./internal/state/; go build; gofmt.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: QUIET

Plan: Enterprise

Run ID: a1a60954-2eac-4c2c-956e-aeada06b6822

📥 Commits

Reviewing files that changed from the base of the PR and between f12acb8 and 0b0d474.

📒 Files selected for processing (3)

• controllers/clusterpolicy_controller.go
• internal/state/manager.go
• internal/state/metrics.go

Included review availability: Your plan provides up to 12 included reviews per hour; 11 remain after this review.

---

📝 Walkthrough

Walkthrough

The change adds the StateDurationSeconds Prometheus histogram. The metric uses controller and state labels and is registered with the controller-runtime metrics registry. stateManager stores the CRD kind and records synchronization duration. ClusterPolicy reconciliation records duration for each reconciliation state.

Merge Risk: ⚪ Minimal · up to 0b0d4

The change adds per-state reconciliation duration metrics without any identified merge-blocking risk; it is merge-ready after normal checks and review.

---

_{Comment @coderabbitai help to get the list of available commands.}