17. Rate Limits#

The platform may limit request frequency (RPM / concurrency). When a limit is triggered, the API returns 429 Too Many Requests.

Recommendations:

• On 429, retry with exponential backoff (e.g. 1s, 2s, 4s, ...).
• Keep client concurrency under control to avoid sudden bursts.
• Submit batch jobs off-peak or serially where possible.