Rate limits

Requests are limited per partner key and per user. Defaults are 6,000 requests per minute per partner key and 300 per minute per user; deployment configuration can change those values. Inspect the response's rate-limit headers and honor Retry-After on 429.

Streams also have per-user and per-replica socket limits; reaching them returns 429 stream_capacity. Reconnect with backoff. Request limits, execution capacity and user spend budget are separate constraints. Read the error's reason to distinguish them.