Qdrant Multi-Node Cluster

Scalable vector database deployment for efficient similarity search across multiple nodes

📖 Overview

This project demonstrates a scalable, distributed deployment of Qdrant, a high-performance vector database. It showcases how to set up multiple Qdrant nodes in a clustered configuration, enabling efficient vector search operations with high availability and performance.

Qdrant is designed for enterprise-grade vector similarity search, supporting a wide range of use cases:

• Semantic text search: Find documents with similar meaning, not just keywords
• Image similarity: Locate visually similar images
• Recommendation systems: Suggest products, content, or services
• Anomaly detection: Identify outliers in vector spaces
• Chatbot knowledge base: Power semantic retrieval for AI assistants

✨ Key Features

• 🔄 Scalable Multi-Node Architecture: Deploy 3+ Qdrant nodes that work as a unified cluster
• 📊 Dynamic Sharding: Distribute vector data across nodes with customizable sharding strategies
• 🏠 High Availability: Built-in replication for fault tolerance and continuous operation
• 📈 Monitoring Stack: Integrated Prometheus and Grafana for real-time metrics visualization
• 🐍 Python Client Integration: Comprehensive demo application showing cluster interaction
• 🐳 Docker-Based Deployment: Simple setup using Docker Compose
• 🔧 Detailed Configuration: Extensive options for tuning and optimizing performance

🚀 Quick Start

Prerequisites

• Docker and Docker Compose
• Python 3.8+
• Git

Installation

# Clone the repository
git clone https://github.com/Mohitkr95/qdrant-multi-node-cluster.git
cd qdrant-multi-node-cluster
Install the package and dependencies
pip install -e .

Deploy the Cluster

# Start the Qdrant cluster with Prometheus and Grafana
cd deployments/docker
docker-compose up -d

Run the Demo

# Run the demonstration
python src/run_demo.py
Or with custom parameters
python src/run_demo.py --host localhost --port 6333 --points 2000

Access Services

• Qdrant API: http://localhost:6333
• Prometheus: http://localhost:9090
• Grafana: http://localhost:3000 (default login: admin/admin)

📋 Documentation

Comprehensive documentation is available in the docs directory:

• Getting Started Guide
• Architecture Overview
• Configuration Guide
• API Reference
• Performance Tuning
• Troubleshooting

🛠️ Project Structure

qdrant-multi-node-cluster/
├── config/                    # Configuration files
│   ├── grafana.json           # Grafana dashboard configuration
│   └── prometheus.yml         # Prometheus configuration
├── deployments/               # Deployment files
│   └── docker/                # Docker-related files
│       └── docker-compose.yml # Docker Compose configuration
├── docs/                      # Documentation
│   ├── api/                   # API documentation
│   ├── guides/                # User guides
│   └── images/                # Documentation images
├── src/                       # Source code
│   ├── qdrant_demo/           # Main package
│   │   ├── config/            # Configuration settings
│   │   ├── core/              # Core functionality
│   │   └── utils/             # Utility functions
│   └── run_demo.py            # Main entry point
├── tests/                     # Test files
├── LICENSE                    # MIT License
├── Makefile                   # Development commands
├── README.md                  # Project overview
├── requirements.txt           # Python dependencies
└── setup.py                   # Package setup file

📊 Monitoring and Visualization

This project integrates Prometheus for metrics collection and Grafana for visualization, providing real-time insights into your Qdrant cluster's performance.

🔍 Advanced Configuration

Sharding Configuration

Customize sharding to distribute data efficiently:

# In settings.py
SHARD_NUMBER = 4  # Default shard count

Vector Parameters

Configure vector dimensions and distance metrics:

# In cluster_demo.py
client.create_collection(
    collectionname=self.collectionname,
    vectors_config=models.VectorParams(
        size=self.vector_size,  # 768 by default 
        distance=models.Distance.COSINE
    ),
    # ...other parameters
)

Adding More Nodes

Extend the cluster by adding more nodes in docker-compose.yml:

qdrant_node4:
  image: qdrant/qdrant:v1.6.1
  volumes:
    - ./data/node4:/qdrant/storage
  depends_on:
    - qdrant_node1
  environment:
    QDRANTCLUSTERENABLED: "true"
  command: "./qdrant --bootstrap http://qdrantnode1:6335 --uri http://qdrantnode4:6335"

🧪 Testing

Run the test suite:

# Run all tests
make test
Or directly with Python
python -m unittest discover -s tests

🤝 Contributing

Contributions are welcome! See our Contributing Guide for details on how to get started.

• Fork the repository
• Create a feature branch (git checkout -b feature/amazing-feature)
• Commit your changes (git commit -m 'Add amazing feature')
• Push to the branch (git push origin feature/amazing-feature)
• Open a Pull Request

📜 License

This project is licensed under the MIT License - see the LICENSE file for details.

👏 Acknowledgments

• Qdrant for the excellent vector database
• Prometheus and Grafana for monitoring capabilities
• All contributors who help improve this project

📧 Contact

Mohit Kumar - @Mohitkr95

Project Link: https://github.com/Mohitkr95/qdrant-multi-node-cluster