Idempotent API Design and Distributed Rate Limiting with Token Bucket, Leaky Bucket, Nginx, and Redis+Lua

Interface idempotency ensures that repeated requests produce the same result, typically by using a unique business identifier or a token mechanism to lock updates and inserts.

Distributed rate limiting controls traffic across multiple dimensions such as QPS, connection count, transmission speed, black‑white lists, and the whole distributed environment.

Common algorithms include the token‑bucket and leaky‑bucket methods, each with distinct characteristics for handling burst traffic.

Typical implementations:

Guava RateLimiter client‑side limiting (Maven dependency and controller examples).

Nginx limit_req and limit_conn directives with zone, burst, and nodelay settings.

Redis + Lua scripts for atomic limit checks, integrated with Spring Boot via StringRedisTemplate and RedisScript .

Additional components such as a custom AccessLimiter service, Spring configuration, and an AOP annotation are provided to encapsulate the limiting logic and apply it declaratively.