Monitoring Stack

A self-hosted system monitoring stack. The agent (FastAPI + psutil) runs natively on each host (macOS, Raspberry Pi). The server (Flask + SQLite) runs in Docker and is published to Docker Hub as midnightappcoder/statix.

Features

• Live charts — CPU %, RAM %, Disk I/O, Network I/O with configurable timeframes
• Host detail cards — CPU info & temperature, memory, storage, system info, network, uptime, Docker
• OS update check — shows pending updates on macOS (softwareupdate) and Debian/RPi (apt), cached 24 h
• Docker image update check — per-container update badge, cached 24 h
• All active network interfaces — shows every up, non-loopback interface with IP, speed, and link type
• CPU temperature — psutil + sysfs fallback on Linux/Raspberry Pi; osx-cpu-temp on macOS Intel; hidden on Apple Silicon
• Version mismatch warnings — dashboard shows a yellow banner if the server or any client is running a different git SHA than expected; updated automatically on every deploy
• User-configurable alert thresholds — per-card color (Amber/Red/Blue/Green/Purple/custom hex), percentage input, on/off; persisted in localStorage
• Drag-to-reorder & hide cards and charts; order persisted in localStorage
• Multi-host — auto-cycles hosts or pin to a specific machine
• Dark/light theme toggle
• Multi-arch Docker image — linux/amd64 + linux/arm64 (Raspberry Pi)

Quick Usage

1. Start the monitoring server

Requires Docker. Pulls the pre-built image from Docker Hub:

bash <(curl -fsSL https://raw.githubusercontent.com/chandanankush/statix/main/server/install.sh)

Or manually with Docker:

docker run -d \
  --name statix \
  --restart unless-stopped \
  -p 5050:5000 \
  -v statix_data:/app/data \
  midnightappcoder/statix:latest

Or with Docker Compose:

docker compose up -d

Visit http://localhost:5050/dashboard.

2. Install the agent on each host (macOS or Raspberry Pi)

bash <(curl -fsSL https://raw.githubusercontent.com/chandanankush/statix/main/client/install.sh)

Pass flags to skip interactive prompts:

curl -fsSL https://raw.githubusercontent.com/chandanankush/statix/main/client/install.sh \
  | bash -s -- --server-url http://YOUR_SERVER:5050 --interval 30

Re-running the same command upgrades the package and restarts services without losing your config.

To uninstall:

bash <(curl -fsSL https://raw.githubusercontent.com/chandanankush/statix/main/client/uninstall.sh)

3. View metrics

Open http://YOUR_SERVER:5050/dashboard and filter by hostname and timeframe.

Optional: CPU temperature on macOS Intel

brew install osx-cpu-temp

Temperature is shown automatically once the tool is on PATH. Not available on Apple Silicon.

Components

• client/ — FastAPI stats service, forwarder, and one-line installer/uninstaller. See client/README.md.
• server/ — Flask ingestion API, SQLite storage, and Chart.js dashboard. Published as midnightappcoder/statix on Docker Hub. See server/README.md.
• ARCHITECTURE.md — High-level design and data flow.
• docker-compose.yml — Container orchestration for the monitoring server.

Getting Help

• client/README.md — installation, configuration, service management, and API reference for the agent.
• server/README.md — server configuration, Docker usage, API endpoints, and Docker Hub CI setup.
• ARCHITECTURE.md — high-level design and data flow.
• CONTRIBUTING.md — how to set up a dev environment and make changes.
• AGENTS.md — rules and conventions for AI coding agents working in this repo.