Zendesk REST APIs currently respond with a 429 error code after an account has reached its rate limit. The response includes a retry-after header that allows a service to throttle its own requests and comply with standard rate limits. To better protect our quality of service, requests from individual IP addresses that receive 429 error codes and continue to make excessive requests beyond the rate limit within 10 seconds will receive cached 429 responses for the next 10 seconds before being allowed to make requests again. This measure will be introduced on March 11, 2019, to prevent overloading our servers with what could be considered abusive traffic.
If an individual client makes a number of requests matching the following criteria, the client will receive cached 429 responses for 10 seconds.
- Requests to all /api/v2/* endpoints
- Requests that have already reached X-Rate-Limit-Remaining=0
- Requests made after excessive HTTP 429 responses have been given within 10 seconds
Product areas affected:
- Zendesk Support
- Zendesk Guide
- Zendesk Talk
- Channel framework
- Answer Bot
Responses that are rate limited will automatically receive a response with the following message:
"detail":"Rate limited by Zendesk edge protection"
Some strategies to avoid this type of limiting include using the retry-after header data to put a request on hold until the number of suggested seconds passes. Also, we recommend instrumenting your own outgoing traffic to make sure your services are complying with rate limits.
If you have any questions related to scripts or services that you feel might be affected, please contact Zendesk Support for assistance.