Design a Distributed Rate Limiter
Company: Pinterest
Role: Software Engineer
Category: System Design
Difficulty: medium
Interview Round: Technical Screen
Design a rate limiting service for a public API platform. The system must enforce configurable quotas per API key, user, IP address, and endpoint. It should support limits such as requests per second, requests per minute, daily quotas, and short burst allowance. Multiple application servers and API gateways will call the limiter concurrently. The limiter should return allow or deny decisions with a retry-after value, add minimal latency to the request path, scale to high traffic, and remain reliable during partial failures. Discuss algorithms, storage, distributed consistency, failure handling, observability, and trade-offs.
Quick Answer: This question evaluates a candidate's distributed systems and API infrastructure design skills, specifically covering rate-limiting algorithms, concurrency control, storage and consistency trade-offs, scalability, fault tolerance, and observability.