A lightweight, self-hosted Java monitoring agent for collecting system metrics and filtered process statistics, stored directly in your own database.

This project is designed to be simple, transparent, and extensible, without any dependency on third-party SaaS platforms.

✨ Features

• 📊 System metrics
  • CPU usage
  • Memory usage
• 🧩 Process monitoring
  • Track only selected processes (e.g. java, php)
  • Capture CPU, memory, state, user, command, and start time
• ⏱ Configurable interval
  • Run every N seconds (default: 5)
• 🗄 Database-backed storage
  • MySQL / MariaDB
  • Schema managed via Liquibase migrations
• 🔐 Secure local configuration
  • Credentials stored outside the repository
• ⚙️ CLI-based setup
  • One-time setup command
  • Normal start without setup
• 🧵 Non-blocking scheduler
  • Runs continuously in background threads
• 🖥 Cross-platform
  • Linux, macOS, Windows (Java-supported environments)

🧠 Why this project?

Most monitoring solutions today are:

• tightly coupled to SaaS platforms
• heavy on resources
• hard to customize at process level
• opaque in how data is collected

This project focuses on:

• full control of your data
• simple architecture
• process-level visibility
• self-hosted monitoring

It’s suitable for:

• learning how monitoring agents work internally
• building custom dashboards
• controlled or air-gapped environments
• experimentation and extensions

🛠 Tech Stack

• Java
• Maven
• OSHI (system & process metrics)
• MySQL / MariaDB
• Liquibase
• Apache Commons CLI

Prerequisites

• Java 17+
• Maven 3.8+
• MySQL / MariaDB

---

Build the project

mvn clean package

Steps to install on server

Run setup only once on a machine.
This will:

• validate database credentials
• create database tables using Liquibase
• save configuration locally
• configure polling interval
• configure process filters

java -jar agent.jar setup \
  -host localhost \
  -port 3306 \
  -db monitoring \
  -u root \
  -p 'password' \
  -interval 20 \
  -f java,php

⚙️ Setup Options

• -host -> Database host
• -port -> Database port
• -db -> Database name
• -u -> Database user
• -p -> Database password
• -interval -> Polling interval in seconds
• -f -> Process filter (comma-separated list)

Example filters

-f java
-f java,php
-f nginx,mysqld

Only processes whose name or command line matches the filter will be tracked.

🧵 Running in Background

Linux / macOS

nohup java -jar agent.jar > agent.log 2>&1 &

Windows (PowerShell)

Start-Process java -ArgumentList "-jar agent.jar"

Production (Recommended) Run the agent as:

• a systemd service on Linux
• a Windows service

This enables auto-restart, logging, and startup on boot.

⚠️ Performance Notes

Tracking all processes frequently can generate a large volume of data.

Recommended best practices:

• Always use process filters
• Increase polling interval on busy systems
• Track only long-running or critical processes
• Implement data retention policies

🛣 Roadmap

Planned and potential improvements:

• Disk and network metrics
• JVM-level monitoring
• Log file tracking
• Config reload without restart
• Aggregation and retention jobs
• Alerting and thresholds
• systemd / Windows service installers
• REST / push-based reporting
• Dashboard integrations

🤝 Contributing

Contributions are welcome and appreciated.

You can help by:

• fixing bugs
• improving performance
• adding new collectors
• improving documentation
• discussing architecture and design

Please open an issue or submit a pull request.