uptime-monitor - SKILL.md Agent Skill

name: uptime-monitor description: 24/7 OpenClaw uptime monitor. Every 5min cron ping → writes dead.json if down, uptime.json after 7d (168h) continuous alive streak. Use when setting up persistent monitoring (cron setup, streak tracking, status files).

Uptime Monitor

Silent 24/7 sentinel: Tracks OpenClaw/Gateway alive → dead.json (fail) or uptime.json (7d streak). Files in workspace/uptime/.

Quick Setup (One-Shot)

# Install cron (5min pings)
📊 cron add uptime-5m '{"kind":"every","everyMs":300000}' '{"kind":"systemEvent","text":"UPTIME CHECK 👻"}' --sessionTarget main

# View status/logs
📊 cron list
📊 cron runs uptime-5m

Workflow (Auto on "UPTIME CHECK")

Ping: 📊 session_status + openclaw gateway status (via exec).
Success: Update uptime/streak.json (hours += 5/60). If >=168h → write uptime/uptime.json.
Fail: Write uptime/dead.json {ts, downtime_start: now}.
Dirs: Auto-mkdir uptime/.

Streak Reset: On fail → streak=0.

Files (Workspace/uptime/)

streak.json: {"streak_hours": 24.5, "last_ping": 1738746800000}
uptime.json: {"streak_hours": 168.1, "verified": true, "end_ts": 1738746800000} (7d+)
dead.json: {"ts": 1738746800000, "downtime_start": 1738746800000}

Edge Handling

First run: streak=0.
Cron miss: Streak holds (no false-dead).
Manual: message "UPTIME CHECK 👻" triggers.

No alerts/deps. Pure files. Prod eternal.

Script (Optional Exec)

scripts/uptime-check.js: Standalone Node ping (for manual/cron spawn).

Prod: Cron → silent forever.