FastAPI with PostgreSQL and Redis

Simple, production-ready FastAPI service with async PostgreSQL (SQLAlchemy 2 + asyncpg), Redis cache-aside, and structured JSON logging with per-request correlation IDs.

Deploy on Railway

1. Click the button above (or deploy this repo into a new Railway project).
2. Add the PostgreSQL and Redis plugins — Railway injects DATABASE_URL and REDIS_URL automatically; no extra config needed.
3. The app reads $PORT and those URLs on boot. Health check path is /ready (see railway.json), so deploys wait until Postgres and Redis are reachable.

Features

• Async stack: FastAPI + SQLAlchemy 2.0 async + asyncpg.
• Redis cache-aside on item reads; a Redis outage degrades gracefully (falls back to DB).
• JSON logging to stdout with request_id, method, path, status, and duration.
• Health probes: /health (liveness) and /ready (checks Postgres + Redis).
• Railway-friendly: reads DATABASE_URL / REDIS_URL, binds to $PORT, normalizes postgres:// URLs.

Endpoints

Method	Path	Description
GET	/health	Liveness (process up)
GET	/ready	Readiness (Postgres + Redis up)
POST	/items	Create item
GET	/items	List items (limit, offset)
GET	/items/{id}	Get item (cached)
DELETE	/items/{id}	Delete item (invalidates cache)

Interactive docs at /docs.

Run locally

python -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
cp .env.example .env   # adjust DATABASE_URL / REDIS_URL
uvicorn app.main:app --reload

Requires a reachable PostgreSQL and Redis (set their URLs in .env).

Quick test

curl -X POST localhost:8000/items -H 'content-type: application/json' \
  -d '{"name":"hello","description":"world"}'

curl localhost:8000/items/1   # second call is served from Redis
curl localhost:8000/ready

Tests

pip install -r requirements.txt -r requirements-dev.txt
pytest

Tests use in-memory SQLite and a fake Redis, so no live Postgres/Redis is needed.

Notes

• init_db() runs create_all on startup for convenience. For real schema changes, use Alembic migrations instead of relying on auto-create.
• Logs are single-line JSON on stdout; pass structured fields via logger.info("msg", extra={...}).