Skip to main content

Rate limits

The order engine surfaces do not apply a per-client request quota today, and do not answer 429 for volume. Capacity is managed by the platform and agreed with each tenant.

Design integrations as if a quota existed, so they keep working when one is introduced:

  • Treat 429 as retryable: wait for the Retry-After header when present, otherwise back off exponentially with jitter.
  • Retry 500 and 503 the same way, always with the same Idempotency-Key.
  • Poll the outbox at a steady interval rather than in a tight loop; when a poll returns a full page, poll again immediately until it returns less than a page.
  • Batch reads where the API offers it (balances at trip level rather than per line; the trip view rather than separate calls).
Not available

Published per-client quotas and rate-limit response headers are not available. See Availability.