Skip to main content
Relay limits requests per API key with a token bucket. Each key has a bucket that refills at a steady rate; each request takes one token. A full bucket lets you send a short burst above the steady rate. Test keys have a tenth of these limits.
Token bucket for one API key on POST /v1/events: 300 requests arrive at once against a full 200-token bucket; 200 are accepted with 202 and 100 are rejected with 429 Too Many Requests and Retry-After: 1, with nothing stored. One second later the bucket has refilled 100 tokens and the same 100 requests are resent with the same Idempotency-Keys and accepted. Every response carries RateLimit-Limit, RateLimit-Remaining and RateLimit-Reset.

A full bucket takes a burst of 200 events. The rest get 429 and go through a second later, once the bucket has refilled.

​
Headers

Every response says where you are:

​
When you hit the limit

An empty bucket answers 429 Too Many Requests with a Retry-After header in seconds. Nothing is stored. Wait that long, then send the request again with the same Idempotency-Key. The SDK does this for you, up to three times. If you regularly send more than 100 events per second, use POST /v1/events/batch: one request carries up to 100 events and takes one token.

​
Delivery to your endpoints

Rate limits apply to requests you send Relay. Deliveries to your endpoints are limited separately: by default, Relay keeps at most 10 requests in flight to each endpoint. Change it per endpoint in the dashboard if your server can take more, or less.