Rate limiting

Capping how many requests a caller may make in a period, protecting shared capacity from any one client.

Limits are usually expressed per second and per month, and the two fail differently: bursting past a per-second limit is a retry-with-backoff problem, while exhausting a monthly quota is a capacity-planning one. Handling the 429 response deliberately — with backoff and jitter, not an immediate retry — is what separates a resilient client from one that makes an incident worse.