Implement Robust Distributed Rate Limiting with Token Bucket

Design a distributed rate limiting middleware for a high-traffic microservices architecture using TypeScript. Create a token bucket algorithm implementation that works across multiple service instances, supports configurable rate limits per endpoint, and integrates with Redis for shared state tracking. Include type-safe configuration, graceful degradation mechanisms, and clear logging for exceeded rate limit scenarios. Provide concrete implementation strategies for horizontal scaling and preventing potential race conditions.