system-status
system-status
Hermes-style ops agent:监控 Ubuntu / Docker / 公网 IP,告警推飞书。
功能
🐳 Docker 监控:容器状态、重启次数、CPU/内存、健康检查
🖥️ 系统监控:CPU、内存、磁盘、负载
🌐 公网 IP 漂移检测:多源校验,变化告警
💬 飞书告警:富文本卡片 + @指定人
📊 SQLite 历史数据:每次采样入库,默认保留 7 天
🖼️ Web 面板:实时仪表盘 + 趋势图 + 告警流
打开
http://localhost:8888/web/即可
🌐 HTTP / MCP 接口:给 Hermes / 客户端调用
GET /healthGET /api/docker/containersPOST /api/docker/restartGET /api/systemGET /api/ipGET /api/dashboardGET /api/history/system?minutes=60GET /api/history/ipGET /api/alerts/recentGET /mcp/tools
Related MCP server: 6amdev-server-mcp
快速开始
1. 创建飞书机器人
群聊 → 设置 → 群机器人 → 添加 → 自定义机器人 → 拿到 Webhook 和 Secret
2. 配置环境变量
cp .env.example .env
# 编辑 .env,填入飞书 webhook 和 secret3. 启动
docker compose up -d4. 验证
curl http://localhost:8888/health
curl http://localhost:8888/api/docker/containers
curl http://localhost:8888/api/system
curl http://localhost:8888/api/ip配置说明
.env:
变量 | 必填 | 说明 |
| ✅ | 飞书机器人 Webhook |
| ⭕ | 签名校验密钥 |
| ⭕ | @ 的手机号,逗号分隔 |
| ⭕ | 是否 @所有人 |
| ⭕ | 检查间隔(秒),默认 60 |
| ⭕ | HTTP 端口,默认 8888(避开 cadence-rocky:final 已用的 8080) |
接入 Hermes
Hermes 可以直接通过 HTTP MCP 调这些端点。 在 Hermes 配置里加:
mcp_servers:
- name: system-status
type: http
base_url: http://localhost:8888
tools:
- name: list_containers
method: GET
path: /api/docker/containers
- name: system_health
method: GET
path: /api/system
- name: check_public_ip
method: GET
path: /api/ip
- name: restart_container
method: POST
path: /api/docker/restart
body_schema:
type: object
properties:
name: { type: string }然后你可以对 Hermes 说:
"列出所有退出的容器" "重启 mysql-01" "现在的公网 IP 是什么?" "系统健康度怎么样?"
项目结构
system-status/
├── src/
│ ├── index.ts # 入口
│ ├── config.ts # 配置加载
│ ├── mcp/server.ts # HTTP / MCP 接口
│ ├── monitor/
│ │ ├── docker.ts # Docker 监控
│ │ ├── system.ts # 系统监控
│ │ ├── ip.ts # 公网 IP 检测
│ │ └── loop.ts # 监控主循环
│ └── alert/
│ ├── feishu.ts # 飞书 Webhook + 签名
│ └── templates.ts # 告警模板
├── config/rules.yaml # 告警规则
├── Dockerfile
├── docker-compose.yml
└── .env.example告警冷却
同类告警默认 5 分钟 内不重复推(避免告警风暴)。
可在 src/monitor/loop.ts 调整 COOLDOWN_MS。
安全提示
⚠️
docker.sock等于 root 权限,务必只读:ro⚠️ 容器开了
privileged: true,生产环境请评估风险⚠️ 飞书 Webhook 不要提交到 Git
Related MCP Connectors
MCP server for Uptrack uptime monitoring. Manage monitors and incidents from AI agents.
Free uptime monitoring: HTTP/TCP/TLS/DNS + MCP server checks, cron heartbeats, status pages, alerts.
Uptime monitoring with 127 tools across 23 protocols. Tag filtering + Code Mode.
Uptime, SSL, DNS and domain monitoring you can talk to from Claude or any MCP client.
Related MCP Servers
- AlicenseNot gradedqualityFmaintenanceMulti-machine system monitor with a built-in MCP server that enables AI agents to query health metrics, manage processes, schedule cron jobs, and run diagnostics across local and remote machines.133 npmApache 2.0

6amdev-server-mcpofficial
AlicenseBqualityDmaintenanceMCP server for managing Ubuntu servers with AI integration, enabling system monitoring, Docker control, Gitea repository management, Ollama LLM interaction, workspace file operations, and Claude Code assistant via natural language.34MIT- AlicenseAqualityDmaintenanceEnables AI agents to monitor Docker containers, system resources, ports, and logs in real-time.5MIT
- AlicenseNot gradedqualityDmaintenanceOfficial Hyperping MCP server for uptime, API, cron and server monitoring. 26 tools covering monitors, outages and timelines, uptime, response time, MTTR and MTTA, on-call schedules and escalation policies, over a remote Streamable HTTP endpoint with Bearer token auth and no install.1MIT