API Cache Cemo

A small FastAPI backend demonstrating a cache-aside pattern using Redis and PostgreSQL. The API uses Redis as an optional in-memory cache with TTL and falls back to the database as the main data source. This project is intentionally minimal and was focused on backend architecture and tool/library familiarity rather than frontend work.

---

Stack

• Python
• FastAPI
• Redis
• PostgreSQL
• SQLAlchemy

---

How it works

• GET /items/{id}
  • Checks Redis first, falls back to check Postgres on cache miss, then writes to cache with TTL.
• PUT /items/{id}
  • Writes to Postgres and invalidates the related Redis cache key.
• Redis is optional - if unavailable, the API still serves data from the database.

---

Running Locally

python -m venv apiproj
apiproj\Scripts\activate
pip install -r requirements.txt

Create .env file

DATABASE_URL=postgresql+psycopg2://appuser:apppass@localhost:5432/appdb
REDIS_HOST=localhost
REDIS_PORT=6379
CACHE_TTL_SECONDS=60

Start Redis (WSL or local install), then run

uvicorn app:app --reload

Open Link

http://127.0.0.1:8000/ auto redirects to docs

---

Endpoints

• GET /items/{id}
• PUT /items/{id}
• DELETE /items/{id}
• GET /health

---

Notes

• Cached values are JSON-serialized
• Cache keys use the format item:{id}
• TTL is configurable via environment variables