- Rate is the sustained number of requests allowed per second.
- Burst is the maximum number of requests that can be sent in a short spike before throttling kicks in. Once the burst is exhausted, requests are throttled to the sustained rate.
Default limits
| Metric | Limit |
|---|---|
| Rate | 10 requests/second |
| Burst | 10 requests |
Extended limits
For higher-volume workloads, the extended plan increases both limits:| Metric | Limit |
|---|---|
| Rate | 1,000 requests/second |
| Burst | 1,000 requests |
Rate limit responses
When you exceed the rate limit, the API returns a429 Too Many Requests response.
Handling rate limits
When you receive a429 response, wait briefly before retrying. A simple approach:
- Wait 1 second after receiving a
429response. - Retry the request.
- If you receive another
429, double the wait time (2 seconds, then 4 seconds, and so on). - After 5 retries, stop and log the failure for investigation.