health-check

name	health-check
description	Full system health report in plain English with traffic-light ratings. Triggers: "is it working", "health check", "full status", "how's the system".

Health Check

MANDATORY: Run all 7 checks. Present results in plain English with traffic lights.

Checks

Run via SSH. All commands are read-only (GREEN). Run in parallel where possible.

1. Gateway Health

ssh -i ~/.ssh/id_ed25519_github root@72.62.213.231 "openclaw gateway health"

GREEN if responding within 5 seconds. YELLOW if 5-15s. RED if timeout or error.

2. Service Status

ssh -i ~/.ssh/id_ed25519_github root@72.62.213.231 "systemctl is-active openclaw-gateway && systemctl is-enabled openclaw-gateway"

GREEN if active + enabled. RED if not active. YELLOW if active but not enabled (won't start on reboot).

3. Agent Workspaces

ssh -i ~/.ssh/id_ed25519_github root@72.62.213.231 "ls ~/.openclaw/workspace/ 2>/dev/null || echo 'No workspace found'"

Verify the expected files are present (AGENTS.md, SOUL.md, IDENTITY.md, skills/, etc.).

4. Cron Jobs

ssh -i ~/.ssh/id_ed25519_github root@72.62.213.231 "openclaw cron list 2>/dev/null || echo 'Cannot list cron jobs'"

GREEN if expected jobs present and healthy. YELLOW if any in error state. RED if cron list fails.

5. Disk Space

ssh -i ~/.ssh/id_ed25519_github root@72.62.213.231 "df -h / | tail -1"

GREEN if under 80% used. YELLOW if 80-90%. RED if over 90%.

6. Unexpected Processes

ssh -i ~/.ssh/id_ed25519_github root@72.62.213.231 "ss -tlnp | grep -v 'sshd\|openclaw\|systemd\|node'"

GREEN if nothing unexpected. RED if any unexpected port is listening (potential security incident from the autonomous script class of bugs).

7. Recent Errors

ssh -i ~/.ssh/id_ed25519_github root@72.62.213.231 "journalctl -u openclaw-gateway --since '1 hour ago' --no-pager -p err 2>/dev/null | head -20"

GREEN if no errors. YELLOW if warnings only. RED if errors present.

Output Format

=== System Health ===

Overall: GREEN | YELLOW | RED

| Check | Status | Plain English |
|-------|--------|--------------|
| Gateway | [color] | [e.g., "Running and responding in 120ms"] |
| Service | [color] | [e.g., "Active and will start on reboot"] |
| Workspace | [color] | [e.g., "All expected files present"] |
| Cron Jobs | [color] | [e.g., "7 jobs registered, all healthy"] |
| Disk | [color] | [e.g., "Using 45% of 50GB"] |
| Security | [color] | [e.g., "No unexpected network services"] |
| Recent Errors | [color] | [e.g., "Clean — no errors in the last hour"] |

[If any YELLOW/RED: plain-English explanation of what needs attention and what to do]

Rules

• Every check gets a traffic light. No ambiguity.
• Unexpected listening ports are always RED (reference the autonomous HTTP server incident).
• If SSH itself fails, report that first: "Can't reach the server. The VPS might be down or the network might be having issues."
• Don't alarm Trent about normal occasional warnings. Distinguish between "this is fine" and "this needs attention."
• If issues found, offer to run /troubleshoot for deeper investigation.