System Health

这是一款无LLM调用的本地系统健康监控工具，通过crontab定时执行。核心监控6项指标：Gateway响应、PM2/daemon状态、Docker健康、磁盘inode、RAM使用率、证书有效期，分临界与警告阈值。典型场景含查询系统健康、配置检查规则、排查告警、调整检查频率或新增检查项。可通过bash命令手动触发，输出为日志文件，临界状态自动推送告警至OpenClaw agent。

ciklopentan

热度 20.18

最近更新 2026-04-06 12:16:48

分类：自动化依赖：代码执行 / 读写文件 / MCP 服务应用场景：DevOps / 自动化脚本 / 报告撰写

安装技能

复制下面的话术发给你的 Agent，它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill： npx secure-skills-cli add system-health

Metadata

name: system-health
description: Local system health monitoring without LLM API calls. Runs a slim OpenClaw-specific health loop (Gateway, PM2/deepseek-daemon, Docker, disk, RAM, daily cert expiry) via bash cron. Use when: (1) user asks "health", "status", "проверь систему", "мониторинг", (2) want to configure or review health checks, (3) investigating CRITICAL/WARNING alerts, (4) adjusting check frequency or adding/removing checks. Replaces LLM-based heartbeats with zero-API-cost local monitoring.

System Health — v3.0 (2026-04-06)

Local system monitoring skill. No LLM calls. Runs via system crontab.

Quick Usage

# Run now
bash ~/.openclaw/workspace/skills/system-health/scripts/health-check.sh

# See latest report
cat ~/.openclaw/workspace/logs/health-latest.md

# See history
tail -20 ~/.openclaw/workspace/logs/health-history.log

Checks

#	Check	Critical if	Warning if
1	Gateway	Not responding on localhost:18789	—
2	PM2 / deepseek-daemon	daemon not online	PM2 missing
3	Docker	daemon unavailable or unhealthy containers	—
4	Disk + inode	≥95% full/inode	≥85% full/inode
5	RAM	≥90% used	≥80% used
6	Cert expiry	<7 days remaining (daily gate at 03:xx)	—

Cron Setup

# Every 5 minutes
(crontab -l 2>/dev/null | grep -v 'health-check.sh'
echo '*/5 * * * * /usr/bin/env bash /home/irtual/.openclaw/workspace/skills/system-health/scripts/health-check.sh >> /home/irtual/.openclaw/workspace/logs/health-cron.log 2>&1'
) | crontab -

Alerting

When status = CRITICAL, the script calls:

openclaw system event --text "🔴 Health Check CRITICAL: ..." --mode now

This wakes the agent and routes the alert via configured channels.

Adding a Check

Add a check_XXX function in scripts/health-check.sh
Register it in the "Run checks" section
Document in the table above

Pattern:

check_XXX() {
    # logic
    # CRITICAL+=("❌ ...") or WARNINGS+=("⚠️  ...") or OK_ITEMS+=("✅ ...")
}

Files

scripts/health-check.sh — main script (runs from cron)
~/.openclaw/workspace/logs/health-latest.md — latest report (markdown)
~/.openclaw/workspace/logs/health-history.log — append-only history (trimmed to last 1000 lines)
~/.openclaw/workspace/logs/health-cron.log — cron execution output