Build a RAG Chatbot from Scratch — Part 5: Deployment and Production Considerations

FROM python:3.12-slim

WORKDIR /app

COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt

COPY . .

EXPOSE 8000

CMD ["uvicorn", "app.main:app", "--host", "0.0.0.0", "--port", "8000", "--workers", "2"]

# app/middleware/rate_limit.py
from fastapi import Request, HTTPException
from app.core.cache import redis_client
import time

RATE_LIMIT = 20  # requests per minute

async def rate_limit_middleware(request: Request, call_next):
    if request.url.path == "/chat/stream":
        client_ip = request.client.host
        key = f"rate:{client_ip}"
        current = redis_client.get(key)

        if current and int(current) >= RATE_LIMIT:
            raise HTTPException(429, "Too many requests. Please wait.")

        pipe = redis_client.pipeline()
        pipe.incr(key)
        pipe.expire(key, 60)
        pipe.execute()

    return await call_next(request)

version: '3.8'

services:
  api:
    build: .
    ports:
      - "8000:8000"
    environment:
      OPENAI_API_KEY: ${OPENAI_API_KEY}
      CHROMA_PERSIST_DIR: /data/chroma
    volumes:
      - chroma_data:/data/chroma
    healthcheck:
      test: ["CMD", "curl", "-f", "http://localhost:8000/health"]
      interval: 30s
      retries: 3

  redis:
    image: redis:7-alpine
    volumes:
      - redis_data:/data

volumes:
  chroma_data:
  redis_data:

# Track and limit costs
COST_PER_1K_TOKENS = {
    "gpt-4o-mini": {"input": 0.00015, "output": 0.0006},
    "text-embedding-3-small": 0.00002
}

def estimate_cost(model: str, input_tokens: int, output_tokens: int) -> float:
    costs = COST_PER_1K_TOKENS.get(model, {})
    input_cost = input_tokens / 1000 * costs.get("input", 0)
    output_cost = output_tokens / 1000 * costs.get("output", 0)
    return round(input_cost + output_cost, 6)

@app.get("/health")
def health():
    try:
        chroma_client.heartbeat()
        redis_client.ping()
        return {"status": "healthy", "vector_store": "ok", "cache": "ok"}
    except Exception as e:
        return {"status": "unhealthy", "error": str(e)}

docker-compose up -d --build
curl http://localhost:8000/health
# {"status":"healthy","vector_store":"ok","cache":"ok"}