[monitoring] Add comprehensive monitoring for new control plane

- Fix agent kube-rbac-proxy upstream port (4215->4270) to actually
  enable metrics scraping
- Add drbd-reactor sidecar with dedicated kube-rbac-proxy to agent
  DaemonSet for DRBD device metrics (quorum, sync, connections)
- Add second PodMonitor endpoint for DRBD metrics scraping
- Add controller headless Service and ServiceMonitor for Prometheus
  scraping on port 4272
- Add controller-runtime Grafana dashboard with reconcile RPS,
  heatmap, queue metrics, timing/latency, CPU/memory panels
- Add VPA policies for new agent containers

Signed-off-by: Ivan Ogurchenok <ivan.ogurchenok@flant.com>

Author: IvanOgurchenok