Limits and errors

Read every page, respect rate limits and recover safely from errors.

1. Rate limits

   Limits are configurable. GET /v1/operations/capabilities reports requestsPerMinute and authenticationRequestsPerMinutePerIp. After a 429, honour Retry-After and wait longer if throttling continues. Batch reads accept up to 100 IDs; individual endpoints may have smaller page or payload limits.

2. Read every page

   Use the continuation rule for the endpoint you called, keeping the same filters. Lists can change while you read them; use snapshots when you need a consistent copy.

   Pagination	Continue with	Stop when
   Cursor	Pass next as after.	next is null.
   Offset with nextOffset	Pass nextOffset as offset.	nextOffset is null.
   Offset with hasMore	Increase offset by the number of returned items.	hasMore is false.
   Software	Pass the numeric next value as after; restart if updatedAtUtc changes.	next is null.
   Event feed	Pass next as after.	hasMore is false.

3. Handle errors

   Read the error code as well as the HTTP status. If the response includes requestId, quote it when contacting support.

   HTTP status	What to do
   400	Fix the request.
   401	Check credentials or obtain a new access token.
   403	Check scope, permissions and whether access is enabled.
   404	The resource may be missing or outside your scope.
   409	Read the error code and current state before deciding.
   410	The snapshot or event cursor expired. Start a new baseline.

4. If a write times out

   A timeout or 503 does not prove the action failed. For writes that return operation receipts, look up GET /v1/operations/by-key?idempotencyKey={key} before retrying. For maintenance creation, reuse the original Idempotency-Key and body.