Set priorityClassName to system-cluster-critical for frr-k8s to avoid pod eviction

FRR-K8s pods currently run without a cluster-critical priority class, making them vulnerable to eviction under node pressure. Set priorityClassName: system-cluster-critical to ensure routing remains available during resource pressure. This can be configured in MetalLB helm values.

metallb_helm_values:
  frr-k8s:
    frrk8s:
      priorityClassName: "system-cluster-critical"