Rate limits

Awell does not publish request-rate limits for the Orchestration API. There is no documented ceiling on requests per second or per minute, and nothing that says whether one would apply per API key, per tenant, or per environment.

That is a statement about what is documented, not a promise that you can send unlimited traffic. If you are planning something high-volume — a bulk import, a nightly sync over a large population, a backfill — ask support@awellhealth.com what the ceiling is before you build against an assumption.

The limit that does exist

List queries cap the number of records you get back per request:

• count above 100 does not error. The query succeeds and returns 100 records.
• Omitting pagination altogether returns 10.

That is a result-size limit rather than a rate limit, and it is the one most likely to surprise you. See Pagination.

Behave well anyway

Whether or not a rate limit is enforced, these hold:

• Retry 5xx, with exponential backoff. A 5xx is an error inside Awell and may clear on its own. Back off between attempts rather than retrying immediately.
• Do not retry a 200. GraphQL returns 200 even when the request failed, so a retry loop that branches on the HTTP status will hammer the API repeating a request that will never succeed. See Errors.
• Page rather than widening. If a list is large, walk it with offset instead of raising count, which is capped anyway.

Related

• Errors
• Pagination