## Request hygiene | Recommendation | Why it matters | | -------------------------------- | -------------------------------------------------------------------- | | Send the smallest useful request | Smaller responses and easier debugging | | Set a reasonable `limit` | Avoids fetching more data than needed | | Follow endpoint pagination | Use `limit`, `offset` or time-range parameters exactly as documented | | Cache slow-changing data | Configuration and reference metadata do not need constant refetching | | Retry with backoff | Avoids retry storms during temporary errors | ## Request guidance | Use case | Recommendation | | ------------------ | -------------------------------------------------------- | | List views | Request only the fields and page size needed by the UI | | Detail views | Fetch by stable resource identifiers | | Background sync | Use pagination and store progress on the client side | | Temporary failures | Retry with exponential backoff and a maximum retry count | ## Pagination flow Include the `limit` that matches your UI or processing batch size. Inspect whether the endpoint returns enough data for the current view or batch. Increase `offset` or advance the time range according to that endpoint's reference page. ## Error handling | Status | Suggested handling | | ------ | --------------------------------------- | | `400` | Fix query parameters or time format | | `401` | Check account credentials | | `404` | Check the requested resource ID or path | | `429` | Slow down and retry later | | `503` | Retry with backoff after a short delay |