We've set API rate limits to protect Katana from API traffic spikes that could put our systems at risk. These limits also help ensure that our platform remains stable and usable for everyone. We evaluate the number of requests sent to an API and limit them if they surpass the allowed amount.
By default, Katana API allows 60 requests per 60 seconds.
Your current rate limit quota is provided in the response headers of each request.
| Header | Description |
|---|---|
| X-Ratelimit-Limit | Number of requests available for this application. |
| X-Ratelimit-Remaining | Number of requests remaining in quota. |
| X-Ratelimit-Reset | The timestamp (as milliseconds since epoch) when the quota will reset. |
Example
# Example with the default rate limit of 30
HTTP/1.1 200 OK
Content-Type: application/json
X-RateLimit-Limit: 30
X-RateLimit-Remaining: 29
X-RateLimit-Reset: 1640995199
# If you make more than the allowed requests
HTTP/1.1 429 Too Many Requests
Content-Type: application/json
Retry-After: 15
If the rate limit is surpassed, you will receive a 429 error code indicating that you have sent too many requests within a period. If this happens, you will receive a 'Retry-After' header with a value indicating how many seconds you need to wait before sending another request.
Handling Rate Limiting: Throttling on the Consumer Side
To avoid hitting the API rate limits, implement throttling in your client application. Throttling controls the rate of outgoing requests, ensuring you do not exceed the allowed quota (60 requests per 60 seconds by default).
A simple approach is to queue your requests and release them at a controlled interval, such as one request per second. Additionally, always check the 429 Too Many Requests response and the Retry-After header. If you receive this response, pause sending new requests for the specified duration before retrying.
By proactively managing your request rate, you can maintain uninterrupted access to the Katana API and avoid errors caused by exceeding rate limits.