Rate limits
The endpoint is protected by request-rate limits applied at the network edge. Requests are counted per client IP address over a rolling 5-minute window.
| Scope | Limit (requests per 5 minutes) |
|---|---|
| Per client IP | 500 |
Requests that exceed a limit are rejected at the edge with HTTP status
403 until the window falls back below the threshold. The limits are
sized for normal interactive and scripted use; well-behaved clients
rarely reach them.
Recommendations
- Retrieve data in a moderate number of larger queries rather than many small ones. Each query may return up to 10 000 rows — see Pagination.
- When iterating over filter windows, run the requests sequentially or with limited concurrency rather than firing them all at once.
- On HTTP
403for a previously valid request, pause briefly and retry; a short back-off keeps a client comfortably under the per-IP limit.
Per-query result limit
Independently of the request-rate limits, a single query is capped at 10 000 rows. Queries whose result set exceeds the cap are rejected rather than truncated. See Pagination for how to split larger result sets into filter windows.