Loady

A lightweight, fast, and extensible load balancer written in Go.

Loady helps you distribute traffic across multiple backend servers with health checks, hot-reloadable configuration, and support for multiple balancing algorithms (round-robin, least-connections, IP hash, etc.).

---

✨ Features

• 🔄 Multiple load balancing algorithms (round-robin, least-connections, IP hash, …)
• ❤️ Health checks with automatic failover
• ⚙️ Hot-reloadable configuration (no restart required)
• 🔐 Graceful shutdown & error handling
• 📊 Metrics & logging for monitoring traffic
• 📦 Cross-platform binaries for Linux, macOS, and Windows

---

🚀 Installation

Using Go

go install github.com/adel-hadadi/loady@latest

From Release

Download a prebuilt binary from the Releases page.

Example for Linux:

curl -L https://github.com/adel-hadadi/loady/releases/latest/download/loady_linux_amd64 -o loady
chmod +x loady
sudo mv loady /usr/local/bin/

---

⚡ Usage

Run Loady with a simple config file:

loady --config ./config.yaml

Note

Loady can also run without that --config flag but by default its looking for config file in /etc/loady/config.yml.

Example config.yaml

algorithm: round-robin
servers:
  - http://127.0.0.1:8081
  - http://127.0.0.1:8082
  - http://127.0.0.1:8083

healthcheck:
  interval: 5s
  timeout: 2s
  path: /health

---

🔀 Supported Algorithms

• Round Robin – evenly distributes requests
• Least Connections – sends new requests to the server with the fewest active connections
• IP Hash – same client IP always goes to the same backend

---

🩺 Health Checks

Loady continuously probes backend servers and automatically removes unhealthy nodes from the pool. When they recover, they’re added back automatically.

Example log:

[INFO] server 127.0.0.1:8082 is unhealthy
[INFO] server 127.0.0.1:8082 is back online

---

📊 Observability with Prometheus

Loady now integrates with Prometheus to provide real-time metrics about your load balancer and the services it manages. This allows you to monitor the performance and health of your applications easily.

Metrics Exported

Some of the key metrics Loady exposes:

Metric	Description
http_request_duration_seconds	Duration of HTTP requests handled by Loady.
http_total_operations_processed	Total number of requests processed by Loady.
healthy_servers_count	Number of servers currently marked as healthy.

How to Use

1. Run Loady with Prometheus metrics enabled (enabled by default on /metrics endpoint).
2. Point your Prometheus server to scrape Loady:

scrape_configs:
  - job_name: 'loady'
    static_configs:
      - targets: [ '<LOADY_HOST>:<METRICS_PORT>' ]

3. Visualize metrics with Grafana or any other Prometheus-compatible dashboard.

Tip

These metrics help you track response times, detect unhealthy servers quickly, and analyze traffic patterns for better load balancing decisions.

---

🛠 Development

Clone the repo and build:

git clone https://github.com/adel-hadadi/loady.git
cd loady
go build -o loady ./cmd

Run tests:

go test ./...

---

📦 Roadmap

• Add sticky sessions
• Prometheus metrics
• gRPC load balancing
• Advanced configuration via API
• More detailed logging

---

🤝 Contributing

Contributions are welcome! Please open an issue or PR to discuss changes.

---

📜 License

MIT License – see LICENSE for details.