Skip to content

README.md

Latest commit

 

History

History
254 lines (191 loc) · 5.4 KB

README.md

File metadata and controls

254 lines (191 loc) · 5.4 KB

Docker Deployment for AWS RAG Application

This directory contains Docker configuration files for deploying the AWS RAG (Retrieval-Augmented Generation) application.

Prerequisites

  • Docker (version 20.10 or later)
  • Docker Compose (version 2.0 or later)
  • AWS credentials configured (via AWS CLI, environment variables, or IAM roles)
  • Pinecone API key and index

Quick Start

  1. Create environment file:

    # From the project root
cp env.example .env
# Edit .env with your actual credentials

  2. Start the application:

    cd deployment/docker
docker-compose up -d

  3. Check the application:

Configuration

Environment Variables

The application requires the following environment variables:

Required

  • PINECONE_API_KEY: Your Pinecone API key
  • AWS_REGION: AWS region (default: us-east-1)

Optional (with defaults)

  • PINECONE_ENVIRONMENT: Pinecone environment (default: us-east-1-aws)
  • PINECONE_INDEX_NAME: Pinecone index name (default: rag-documents)
  • BEDROCK_EMBED_MODEL_ID: Embedding model (default: amazon.titan-embed-text-v2:0)
  • BEDROCK_LLM_MODEL_ID: LLM model (default: us.anthropic.claude-sonnet-4-20250514-v1:0)
  • LOG_LEVEL: Logging level (default: INFO)

AWS Credentials

You can provide AWS credentials in several ways:

  1. Environment variables (in .env file):

    AWS_ACCESS_KEY_ID=your_access_key
AWS_SECRET_ACCESS_KEY=your_secret_key

  2. AWS CLI configuration (recommended):

    aws configure

  3. IAM roles (for EC2/ECS deployment)

  4. Instance profiles (for EC2 deployment)

Files

  • Dockerfile: Multi-stage Docker build configuration
  • docker-compose.yml: Service orchestration configuration
  • README.md: This documentation

Docker Commands

Build and Start

# Build and start services
docker-compose up -d

# Build without cache
docker-compose build --no-cache

# Start with logs
docker-compose up

Management

# Stop services
docker-compose down

# Restart services
docker-compose restart

# View logs
docker-compose logs -f rag-api

# Check status
docker-compose ps

Development

# Start with auto-reload (for development)
docker-compose up -d

# Execute commands in container
docker-compose exec rag-api bash

# View real-time logs
docker-compose logs -f rag-api

Testing the Deployment

After starting the services, you can test the deployment manually:

  1. Check service health:

    curl http://localhost:8000/health

  2. Test a sample query:

    curl -X POST http://localhost:8000/query \
  -H "Content-Type: application/json" \
  -d '{"query": "What are AWS security best practices?"}'

  3. View API documentation:

    open http://localhost:8000/docs

  4. Check container status:

    docker-compose ps

Troubleshooting

Common Issues

  1. Port 8000 already in use

    # Find and kill the process using port 8000
lsof -i :8000
kill -9 <PID>

  2. Docker build fails

    # Clean Docker cache and rebuild
docker system prune -f
docker-compose build --no-cache

  3. Service fails to start

    # Check logs for errors
docker-compose logs rag-api

# Check container status
docker-compose ps

  4. Health check fails

    # Check if the service is responding
curl -v http://localhost:8000/health

# Check application logs
docker-compose logs rag-api

  5. AWS/Pinecone connection issues

    • Verify your credentials in the .env file
    • Check AWS region and Pinecone environment settings
    • Ensure Bedrock models are enabled in your AWS account
    • Verify Pinecone index exists and is accessible

Debug Mode

To run in debug mode with more verbose logging:

  1. Update docker-compose.yml:

    environment:
  - LOG_LEVEL=DEBUG

  2. Restart the service:

    docker-compose restart

Performance Tuning

For production deployment, consider:

  1. Resource limits in docker-compose.yml:

    deploy:
  resources:
    limits:
      memory: 2G
      cpus: '1.0'

  2. Health check intervals:

    healthcheck:
  interval: 60s
  timeout: 30s
  retries: 3

  3. Logging configuration:

    logging:
  driver: "json-file"
  options:
    max-size: "10m"
    max-file: "3"

Security Considerations

  1. Never commit .env files with real credentials
  2. Use secrets management for production (Docker Secrets, AWS Secrets Manager)
  3. Run as non-root user (already configured in Dockerfile)
  4. Use specific image tags instead of latest for production
  5. Regularly update base images for security patches

Production Deployment

For production deployment:

  1. Use a reverse proxy (nginx, traefik)
  2. Enable HTTPS/TLS
  3. Configure proper logging and monitoring
  4. Use container orchestration (Docker Swarm, Kubernetes)
  5. Implement proper backup strategies
  6. Set up CI/CD pipelines

Support

If you encounter issues:

  1. Check the logs: docker-compose logs rag-api
  2. Verify your environment configuration
  3. Ensure all prerequisites are met
  4. Test endpoints manually using curl commands above