Rate limits
Each route has a weight determining how many requests it counts for. Heavier endpoints, and
endpoints operating on multiple symbols, carry a heavier weight. The default weight is 1 unless
stated on the endpoint.
When you receive a 429, back off rather than continuing to send requests.
/api/* and /sapi/* differ — check which family you are calling
/api/* and /sapi/* apply rate limits differently, and this matters for CAAS and OMS Toolkit
because their endpoints are all /sapi/*.
/api/* | /sapi/* | |
|---|---|---|
| How limits apply | IP and UID limits are independent | Each endpoint is marked as IP- or UID-limited, with its own weight |
| Limit | All endpoints share 6,000 per minute per IP | Each IP-limited endpoint has an independent 12,000 per minute; each UID-limited endpoint an independent 180,000 per minute |
| Weight header | X-MBX-USED-WEIGHT-(intervalNum)(intervalLetter) | X-SAPI-USED-IP-WEIGHT-1M or X-SAPI-USED-UID-WEIGHT-1M |
| Order-count header | X-MBX-ORDER-COUNT-(intervalNum)(intervalLetter) on successful order responses, per UID | not applicable |
| Exceeding the limit | 429, escalating to an automated IP ban (418) if you keep sending | always 429 — the 418 IP-ban escalation does not apply |
IP bans (/api/* only)
- Repeatedly violating rate limits, or failing to back off after
429s, results in an automated IP ban (HTTP418). - IP bans are tracked and scale in duration for repeat offenders, from 2 minutes to 3 days.
- A
Retry-Afterheader is sent with418and429responses, giving the number of seconds to wait — to avoid a ban in the case of a429, or until the ban ends in the case of a418. - These limits are based on IP, not on the API key.
此页面是否有帮助？
Last modified on