The API client is making too many concurrent requests, and some are being throttled (429)

plamenj

We are repeatedly getting Error 429 (The API client is making too many concurrent requests,
and some are being throttled.) to screening requests in Pilot. However there have no more than 6 at the time, which is considerably less than the suggested 50 limit. Is there a lower limit set for Pilot environment?

Mehran.Ahmed Khan

@plamenj

"Error 429 Too Many Requests" is sent as a response when the user has exhausted the maximum number of API calls per second. The users are advised to send 5 API requests per second(suggested limit is not 50) and by exceeding this limit they would receive the error 429. The error depends on the maximum number of API calls per second and the available capacity on the WC1 platform (number of API calls being sent to the WC1 server by all the users at that instance).

In order to deal with this error, please retry the failed API calls after a brief delay and have a mechanism at a client's end to detect 429 error and downscale the number of API calls fired accordingly for better results.

Here is an excerpt from the "Quality of Service" section of our WC1 API documentation for your reference.

"The throttling limits are applied to all requests made by a user of the API, so an HTTP 429 response for one method would require a delay waiting period for all other API requests that may be processing during the same time period, for the same API user. Appropriate error handling that can check for errors such as HTTP 429 and retry requests using a policy such as exponential backoff should, therefore, be added to all World-Check One API calls.

Requests made to the World-Check One API are rate limited to ensure a consistent experience for all customers, however exact throttling limits may vary over time to adjust for available capacity on the platform. When making requests to the World-Check One API, it is advisable to check for <strong>HTTP 429 Too Many Requests</strong> response codes and attempt to retry the affected requests after a brief delay."

For more detailed information, I request you to refer the "Quality of Service" section of our documentation.

Hope this answers your question, if you have any further queries feel free to reach out to us!!