/health-check

name: health-check
description: |
  System health check dispatched by the activity monitor via Control queue.
  Checks PM2 services, disk space, and memory usage.
  Use when receiving a control message containing "health-check".
user-invocable: false
allowed-tools: Bash, Read, Grep

System Health Check

Periodic system health check delivered via the C4 Control queue.

When to Use

• Receiving a control message with "health-check" in the content
• The activity monitor enqueues this automatically at regular intervals

Steps

1. Check PM2 Services

node ~/zylos/.claude/skills/health-check/scripts/pm2-snapshot.js

The helper parses `pm2 jlist` internally and prints only the allowlisted health fields (`name`, `status`, `pid`, and `restart_time`). Never run or print raw `pm2 jlist` output because it contains each process's inherited environment. Every service should have `status: "online"`. Record which services are stopped or errored.

2. Check Disk Space

df -h / /home 2>/dev/null || df -h /

Thresholds:

• OK: < 80% used
• Warning: 80-90% used
• Critical: > 90% used

3. Check Memory

free -m

Thresholds:

• OK: < 80% used
• Warning: 80-90% used
• Critical: > 90% used (or swap > 50% used)

4. Report Results

If all checks pass, log to `~/zylos/logs/health.log`:

[YYYY-MM-DD HH:MM:SS] Health Check: PM2 X/X online, Disk XX%, Memory XX% - ALL OK

If any issues found, notify whoever is most likely to help: 1. Check your memory files for a designated owner or ops person 2. If none designated, notify the person you normally work with most 3. Use `c4-send.js` with the appropriate channel and endpoint to send the alert

Issue Resolution

| Issue | Action | |-------|--------| | PM2 service stopped | `pm2 restart <service>` and report | | High disk usage | Check logs directories, report findings | | High memory / swap | Report findings, check for runaway processes |