Futures Rate Limits
REST Request Limits
Request limits are determined from cost associated with each API call and rate limiting budgets depend on which path the endpoint uses. Public endpoints do not have a cost and therefore do not count against any rate limiting budget.
For /derivatives endpoints, clients can spend up to 500 every 10 seconds.
The below table displays the cost associated with each API call for /derivatives endpoints:
| Endpoint | Cost |
|---|---|
| sendorder | 10 |
| editorder | 10 |
| cancelorder | 10 |
| batchorder | 9 + size of batch |
| accounts | 2 |
| openpositions | 2 |
| fills (without lastFillTime specified) | 2 |
| fills (with lastFillTime specified) | 25 |
| cancelallorders | 25 |
| cancelallordersafter | 25 |
| withdrawaltospotwallet | 100 |
| openorders | 2 |
| orders/status | 1 |
| unwindqueue | 200 |
| GET leveragepreferences | 2 |
| PUT leveragepreferences | 10 |
| GET pnlpreferences | 2 |
| PUT pnlpreferences | 10 |
| transfer | 10 |
| transfer/subaccount | 10 |
| subaccount/{subaccountUid}/trading-enabled | 2 |
| self-trade-strategy | 2 |
On the Batch Order endpoint, the cost is 9 + size of the batch of the requests. For example, a batch of 10 order requests, (send, edit, and or cancel) sent through the Batch Order endpoint would cost 19.
If the API limit is exceeded, the API will return error equal to apiLimitExceeded.
For /history endpoints, clients have a pool of up to 100 tokens that continually replenishes at a rate of 100 every 10 minutes.
The below table displays the cost associated with each API call for /history endpoints. The accountlog rate limit cost varies on the optional 'count' parameter (default count is 500):
| Endpoint | Cost |
|---|---|
| historicalorders | 1 |
| historicaltriggers | 1 |
| historicalexecutions | 1 |
| accounglogcsv | 6 |
| accountlog (count: 1 - 25) | 1 |
| accountlog (count: 26 - 50) | 2 |
| accountlog (count: 51 - 1000) | 3 |
| accountlog (count: 1001 - 5000) | 6 |
| accountlog (count: 5001 - 100000) | 10 |
Example
The following shows the return of call of the sendorder endpoint where the API limit has been exceeded.
{
"result": "error",
"serverTime": "2016-02-25T09:45:53.818Z",
"error": "apiLimitExceeded"
}
Websocket Limits
There are limits to both the number of connections a client can have open concurrently and the number of requests an individual connection can make. Limit values are subject to change and additional limits may be added in the future.
The current limits are:
| Resource | Allowance | Replenish Period |
|---|---|---|
| Connections | 100 | N/A |
| Requests | 100 | 1 second |