PROJECT // MICROSERVICE BACKEND
SpringSentinel
A Spring Boot REST backend that combines PostgreSQL persistence with Redis guardrails for authenticated social interactions.
Problem
Concurrent bot interactions, nested comments, authentication, and rate limits need to remain safe while the backend stays stateless.
Solution
SpringSentinel uses JWT-based identity, role-based access control, PostgreSQL for durable data, and atomic Redis operations for rate limits and virality guardrails.
Architecture
Requests pass through rate-limit and JWT filters before reaching Spring Boot controllers and services. PostgreSQL stores users, posts, comments, and bots; Redis holds short-lived counters, cooldowns, queues, and real-time scores.
Technologies and features
- / JWT authentication with USER and ADMIN roles
- / Nested comment threads with server-calculated depth
- / Atomic Redis Lua guardrails for concurrent bot interactions
- / Distributed rate limiting and notification throttling
- / OpenAPI/Swagger documentation and Docker setup
Engineering decisions
Identity is resolved from the JWT rather than trusted from request bodies. Redis reservations are rolled back when a later PostgreSQL write fails, keeping the two stores aligned.
Outcome
The project demonstrates backend architecture for concurrency-sensitive APIs, with a clear separation between persistent storage and real-time state.