Rate limits#

REST#

Every signed REST request counts against your account's budget. All API keys of the account (team) share one budget.

Limit	Default	Meaning
Requests per second	20	Refill rate of a token bucket
Burst	100	Size of the bucket: how many requests you can send at once after a quiet period
Requests per day	100 000	Per UTC day; the count restarts at 00:00 UTC

Anboto can raise these for your account, up to 100 requests per second, a burst of 500 and 500 000 requests per day.

A request over the budget is answered 429 Too Many Requests with a JSON body and a Retry-After header:

HTTP/1.1 429 Too Many Requests
Retry-After: 1
Content-Type: application/json

{"code": 5, "message": "Rate limit exceeded", "reason": "burst", "retry_after_ms": 50}

Field	Meaning
`code`	`5` = `RATE_LIMIT_EXCEEDED`
`reason`	`burst`: the bucket is empty; wait `retry_after_ms` for the next token. `daily_quota`: the day's requests are used up; `retry_after_ms` runs until 00:00 UTC
`retry_after_ms`	How long to wait, in milliseconds
`Retry-After` header	The same wait in whole seconds, rounded up

A request refused because the bucket is empty does not count toward the daily total.

A request whose X-TIMESTAMP is more than X-RECV-WINDOW ms (default 5000) in the past is rejected, currently with a generic server error rather than 401 (see Errors).

WebSocket#

Limit	Default	On breach
Concurrent connections per account	5; Anboto can raise it up to 50	the connection that goes over the limit gets a code-5 error frame (`connection limit reached (5 per account)`) and is closed (1008). Connections opened earlier stay up
Messages per second per connection	10 (burst 20)	code-5 error frame per dropped message; sustained flooding closes the connection (1008)
Subscriptions per connection (all topics)	50	code-5 error frame
Market data subscriptions per connection	20	error frame ("limit reached")

Error code 5 = RATE_LIMIT_EXCEEDED.