Grafana 儀表板

Vibe Prompt

「幫我在 Grafana 建立一個 K8s 叢集監控儀表板：CPU、記憶體、網路、Pod 狀態，4 個面板。」

Docker Compose

services:
  grafana:
    image: grafana/grafana:latest
    ports:
      - "3000:3000"
    environment:
      GF_SECURITY_ADMIN_PASSWORD: admin
    volumes:
      - grafana_data:/var/lib/grafana

常用面板 Query

| 面板 | PromQL | |------|--------| | CPU 總使用 | 100 - (avg(rate(node_cpu_seconds_total{mode="idle"}[5m])) * 100) | | 記憶體使用 | (1 - node_memory_MemAvailable_bytes / node_memory_MemTotal_bytes) * 100 | | 網路流量 | rate(node_network_transmit_bytes_total[5m]) | | Pod 重啟 | sum(rate(kube_pod_container_status_restarts_total[5m])) |

Alerting Rules

groups:
  - name: node-alerts
    rules:
      - alert: HighCPUUsage
        expr: node_load1 > 2
        for: 5m
        labels: { severity: warning }
        annotations:
          summary: "{{ $labels.instance }} CPU 負載過高 (>2)"
      - alert: DiskFull
        expr: (node_filesystem_avail_bytes / node_filesystem_size_bytes) < 0.1
        for: 10m
        labels: { severity: critical }
        annotations:
          summary: "{{ $labels.instance }} 磁碟即將滿 (>90%)"