Architect: Design a Rate Limiter

Problem

Design a per-user, per-endpoint rate limiter for an ASP.NET Core API serving 10,000 RPS across 4 nodes. Cover:

 1. Algorithm choice (fixed window vs sliding window vs token bucket vs leaky bucket) and why.
 2. State storage (in-memory vs Redis vs hybrid) and trade-offs.
 3. Where the limiter runs (middleware vs gateway vs sidecar).
 4. How you handle bursts, clock skew, and Redis failure.
 5. What headers you return to the client.

Write a concise design (~300-600 words). Submission is saved for human review (not auto-scored).