PERF(resources): optimize cpu requests for worker-

Reduced CPU requests to prevent node-agent scheduling failures:
- code-server: 100m → 50m (actual usage ~5m)
- kubernetes-dashboard metrics-scraper: 100m → 50m (actual usage minimal)

This optimization frees up 100m CPU on worker-node-2, allowing
Velero node-agent (30m request) to be scheduled successfully.

Worker-node-2 CPU allocation before: 840m/1000m (84%)
Worker-node-2 CPU allocation after: 730m/1000m (73%)

code-server/helm-values/code-server.yaml

@@ -28,7 +28,7 @@ service:
 # Resource limits
 resources:
   requests:
-    cpu: 100m  # Reduced for worker-node-2 CPU optimization
+    cpu: 50m  # Reduced from 100m for worker-node-2 CPU optimization (actual usage ~5m)
     memory: 1Gi  # Reduced from 4Gi for worker-node-2 optimization
 
 # Security context