- request-level limits such as
429 Rate limited - product usage limits such as pages, conversions, sources, and concurrency
What is clearly implemented today
The product currently tracks or enforces usage around areas such as:- sources
- concurrent runs
- crawled pages
- converted files
- indexed pages
- API calls
429 responses.
Practical client guidance
- keep polling intervals conservative
- implement backoff and jitter on
429 - avoid bursty user-driven traffic directly against the API
- push long-running work through your backend or job layer