log-audit

name	log-audit
description	Audit all OpenClaw log sources — persistence, retention, size, health. Runs log-audit.sh script.
version	1.0.0
author	repo-man
tags	["logs","audit","monitoring","governance"]

log-audit

Invoke

/log-audit

Steps

1. Run the script

/home/node/.openclaw/scripts/log-audit.sh

2. Format dashboard

Script outputs JSON covering all log sources. Format as:

📋 Log Audit — <date>

Gateway Log:
  Today: <size> | Persisted: <count> files (<size>) | Retention: 7 days
  Pruned: <N> old files this run

Sessions:
  <agent>: <files> files, <size>
  ...
  Total: <files> files, <size> | Pruned: <N> old sessions

Config Audit: <lines> entries | Last change: <date>
Cron Health: <total> recent runs, <failures> failures
Delivery Queue: <pending> pending, <failed> failed
Model Notifications: <lines> entries
Repo-Man Log: <lines> lines

Disk: <total>MB total log footprint

⚠️ Warnings:
  - <warning text>

🔧 Actions taken:
  - <action text>

If status is "ok" and no warnings, use a single line:

✅ Log Audit — All clear. <total>MB total. <N> files pruned.

3. Log result

/home/node/.openclaw/scripts/log-event.sh INFO log-audit "status: <ok|warning> warnings: <N>"

What this script does automatically:

• Persists gateway logs from volatile /tmp/ to ~/.openclaw/logs/gateway/
• Prunes gateway logs older than 7 days
• Prunes session files older than 7 days (keeps min 3 per agent)
• Rotates config-audit.jsonl at 1000 lines (keeps last 500)
• Rotates repo-man.log at 500 lines (keeps last 200)
• Checks cron run history for failures
• Checks delivery queue for stuck/failed messages

Notes

• Do NOT re-implement — always use the script
• Run nightly via cron, not on heartbeat
• If delivery queue has failures, investigate with: cat ~/.openclaw/delivery-queue/failed/*.json | jq .
• If cron failures detected, check: cat ~/.openclaw/cron/runs/*.jsonl | jq -r 'select(.status=="error")'

Intent: Observable [I13]. Purpose: [P-TBD].