Case study
Distributed Rate Limiter & API Gateway
An API gateway with four pluggable, Redis-backed rate-limiting algorithms, per-tenant configuration, circuit breakers and JWT auth — also published as a drop-in Spring Boot Starter.
Built a distributed API gateway with Token Bucket, Sliding Window Log, Fixed Window Counter and Sliding Window Counter, using Redis Lua scripts for atomic distributed counting across gateway instances — eliminating race conditions under concurrency.
Used the Strategy Pattern for algorithm selection, with configuration stored in PostgreSQL and cached in Redis, so algorithms can be switched without redeploying — from the FREE tier (10 req/min) to ENTERPRISE (10,000 req/min).
Integrated Resilience4j circuit breakers with configurable failure-rate thresholds, JWT authentication and API-key management, with a fail-open design that keeps APIs available even during Redis outages.
Published as sentinel-spring-boot-starter with auto-configuration, IDE-friendly configuration metadata and a Maven Central–ready build — any Spring Boot app can add rate limiting with a single dependency.