Skip to content

DEPLOYMENT.md

Latest commit

 

History

History
181 lines (132 loc) · 3.23 KB

DEPLOYMENT.md

File metadata and controls

181 lines (132 loc) · 3.23 KB

Deployment

Quick Start (Development)

uvicorn app.main:app --reload --port 8000

Docker

Build and Run

# Build
docker build -t text2sql-agent .

# Run
docker run -p 8000:8000 --env-file .env text2sql-agent

Docker Compose

# Start all services
docker-compose up -d

# View logs
docker-compose logs -f

# Stop
docker-compose down

Build Options

# CPU-only build (smaller image)
docker build --build-arg TORCH_CPU=true -t text2sql-agent .

# Production target
docker build --target production -t text2sql-agent .

# Development target (includes test tools)
docker build --target development -t text2sql-agent .

Production Deployment

Environment Configuration

Create separate configs for each environment:

# Staging
docker-compose -f docker-compose.yml -f docker-compose.staging.yml up -d

# Production
docker-compose -f docker-compose.yml -f docker-compose.prod.yml up -d

# With load balancer
docker-compose -f docker-compose.yml -f docker-compose.prod.yml -f docker-compose.lb.yml up -d

Kubernetes

# Staging
kubectl apply -k deploy/kubernetes/overlays/staging

# Production
kubectl apply -k deploy/kubernetes/overlays/production

Production Checklist

Database

  • Use PostgreSQL or MySQL (not SQLite)
  • Enable connection pooling
  • Set up read replicas for scaling
  • Configure automated backups

Security

  • Set strong SECRET_KEY
  • Enable HTTPS with valid certificates
  • Configure rate limiting
  • Restrict CORS origins
  • Enable authentication

Performance

  • Enable query caching
  • Use 8-bit quantization for memory efficiency
  • Configure horizontal scaling
  • Set up CDN for static assets

Monitoring

  • Enable Prometheus metrics
  • Set up Grafana dashboards
  • Configure alerting rules
  • Enable structured logging

Infrastructure

  • Set up load balancer
  • Configure auto-scaling
  • Plan disaster recovery
  • Document runbooks

Scaling

Horizontal Scaling

The API is stateless and can be scaled horizontally:

# docker-compose.prod.yml
services:
  api:
    deploy:
      replicas: 3

Auto-Scaling (Kubernetes)

# HPA configuration
spec:
  minReplicas: 2
  maxReplicas: 10
  metrics:
    - type: Resource
      resource:
        name: cpu
        targetAverageUtilization: 70

Health Checks

# Liveness
curl http://localhost:8000/monitoring/health/live

# Readiness
curl http://localhost:8000/monitoring/health/ready

# Full health
curl http://localhost:8000/api/v1/health

Monitoring Stack

# Start with monitoring
docker-compose -f docker-compose.yml -f docker-compose.monitoring.yml up -d

# Access UIs
# Grafana:      http://localhost:3000 (admin/admin)
# Prometheus:   http://localhost:9090
# Jaeger:       http://localhost:16686

GitHub Container Registry

Images are published to GHCR on release:

# Pull latest
docker pull ghcr.io/sakeeb91/text2sql-agent:latest

# Pull specific version
docker pull ghcr.io/sakeeb91/text2sql-agent:1.0.0

Rollback

# Docker Compose
docker-compose pull  # Get previous version
docker-compose up -d

# Kubernetes
kubectl rollout undo deployment/text2sql-agent