Supervision Docker Compose Setup

AI-powered system log analysis with Loki, Promtail, and Ollama.

Quick Start

1. Configure - Edit config/supervision.yml to set your Ollama host and ntfy server
2. Deploy - Run docker compose up -d
3. Access - Open http://localhost:7860 in your browser

Configuration

Before deploying, update config/supervision.yml:

ollama:
  host: "http://YOUR_OLLAMA_HOST:11434"  # Update this!
  model: "qwen3:14b"

notifications:
  ntfy:
    server: "https://YOUR_NTFY_SERVER"  # Update this!
    topic: "supervision-alerts"

Services

• Loki - Log aggregation (port 3101)
• Promtail - Log collection from Docker & systemd
• Supervision - AI analysis web UI (port 7860)

Maintenance

# View logs
docker compose logs -f supervision

# Restart
docker compose restart supervision

# Update
docker compose pull
docker compose up -d

# Stop
docker compose down

Using Custom Image

To build locally instead of pulling from registry:

1. Copy Dockerfile to this directory
2. Copy src/ directory to this directory
3. Copy pyproject.toml to this directory
4. Update compose.yml to use build instead of image:

supervision:
  build:
    context: .
    dockerfile: Dockerfile

Data Persistence

• Loki logs: loki-data volume (7 day retention)
• Analysis history: supervision-data volume