Rate Limits

The below rate limits have been implemented to safeguard our system. Rate limiters use a "fixed window" algorithm to discretely refill the request allowance every 5 seconds.

Market makers are eligible for higher rate limits. To apply for increased rates, please contact our support team.

Burst requests for both REST and WebSockets are refreshed every 5 seconds. E.g a trader can send 25x matching requests in a single burst but must wait 5 seconds before any further requests can be sent.

REST

All requests over the REST API are rate limited per IP. If the limit is crossed,429 Too Many Requests response is returned.

WebSocket

To access the above rate limits, all clients must authorize themselves via the public/login route. Otherwise, a reduced rate limit will be applied. Requests exceeding the rate limit will receive the below response:

{
  id: number
  error: {
    code: -32000,
    message: "Rate limit exceeded",
    data: "Retry after ${cooldown} ms"
  }
}

Only via WebSockets, the live remaining rate limits can be checked using the private/getRateLimits route.

// send getRateLimit request
ws.send(JSON.stringify({
  'public/getRateLimits`,
  {},
  id: 1,
});

// example response
{
  "id": 4,
  "result": {
      "remaining_matching": {
        "remainingPoints": 22,
        "msBeforeNext": 4809,
        "consumedPoints": 3
      },
      "remaining_non_matching": {
        "remainingPoints": 98,
        "msBeforeNext": 4809,
        "consumedPoints": 2
      },
      "remaining_connections": {
        "remainingPoints": 29,
        "msBeforeNext": 8881,
        "consumedPoints": 3
      }
}  

Matching vs Non-Matching

The below requests are counted as "matching" requests. Any request outside of this list is counted as "non-matching".

• private/order
• private/cancel
• private/cancel_all