HomeLab Log Analyzer - Zero-Noise, AI-Powered Daily Summaries

A fully-local, fully-automated system that turns your chaotic Docker logs into clean, structured, actionable data... every night... powered by your own local LLM.

No cloud.
No API fees.
No bullshit.
Just pure self-hosted magic

What This Project Does

This tool automatically:

• Collects logs from all (or selected) Docker containers
• Removes noise, health-checks, pings, spam, repeat lines
• Feeds the cleaned logs into your local LLM via LM Studio
• Generates a structured JSON summary of critical issues, warnings, successes, anomalies, and recommendations
• Sends you a daily email report via N8N
• Runs 100% offline, on your own hardware

Think of it as your homelab’s private Network God.

Original Reddit Post - https://www.reddit.com/r/homelab/comments/1p9f82c/this_is_a_good_use_of_an_llm/

Reddit Post of Release - https://www.reddit.com/r/homelab/comments/1pa8d94/i_built_the_full_local_llm_logsummarizer_system/

How It Works

┌─────────────┐      ┌────────────────┐      ┌──────────────┐      ┌──────────┐
│   N8N       │─────▶│  FastAPI App   │─────▶│  LM Studio   │─────▶│  Email   │
│ (Scheduler) │      │ (Log Analyzer) │      │ (Local LLM)  │      │ Delivery │
└─────────────┘      └────────────────┘      └──────────────┘      └──────────┘
      22:00               ↓                        ↓
                     Pull Logs             Summarize in JSON
                     Filter Noise          Return Clean Insights

N8N triggers ➜ Python FastAPI pulls & filters logs ➜ LM Studio analyzes ➜ N8N emails results.

Features at a Glance

• Docker log ingestion (auto or per-container)
• Noise-filtering engine (health checks, ping/pong, heartbeats, empty lines, etc.)
• Local LLM analysis (Qwen, Phi-3, Llama, anything LM Studio supports)
• Structured JSON output
• Daily summary email
• 100% private & offline
• Lightweight — runs on a mini-PC or NUC
• Plug-and-play with N8N automation

Included in This Repository

• log_analyzer.py — Complete FastAPI backend
• n8n_workflow.json — Drop-in N8N daily summary workflow
• docker-compose.yml — One-command deployment
• Dockerfile — Containerized API
• requirements.txt — All dependencies
• quickstart.sh — Setup in under 60 seconds
• Comprehensive documentation:
  • docs/setup.md
  • docs/models.md
  • docs/troubleshooting.md
  • docs/api.md
  • docs/architecture.md

Example Output (JSON)

{
  "critical_issues": ["Container 'db' restarted unexpectedly"],
  "warnings": ["High memory usage in 'plex'"],
  "successes": ["Backup completed in 'nextcloud'"],
  "recommendations": [
    "Investigate container restarts",
    "Review memory allocation"
  ],
  "container_status": {
    "db": "2 restarts detected",
    "plex": "Memory warning"
  },
  "overall_health": "degraded"
}

Installation

git clone https://github.com/WhiskeyCoder/homelab-log-analyzer
cd homelab-log-analyzer
python log_analyzer.py

Why Local LLM?

Because it’s:

• Free (no API bills)
• Fast (sub-second inference on small models)
• Private (no logs leaving your machine)
• Customizable (your prompt, your workflow)

Models supported:

• Qwen
• Phi-3
• Llama
• Anything LM Studio can serve

📨 Example Daily Email

• 🔴 Critical Issues
• ⚠️ Warnings
• ✅ Successes
• 💡 Recommendations
• 📊 Per-container status
• 🟢 Overall homelab health

Beautiful HTML formatting included.

Contribute

PRs welcome.
New noise-filters, better prompts, new N8N flows — bring them on.

License

MIT — use it, break it, improve it, ship it.

If your homelab produces logs, this tool turns them into clarity.

Give your servers a voice.