Rate Limiting

1 request/second per API key with a 10-request burst capacity

Rate Limit Policy

Rate: 1 request/second per API key
Scope: Applies to all /e9v3 endpoints, including reads
Burst Capacity: 10 requests
Algorithm: Token bucket
HTTP Status: 429 Too Many Requests

The 429 response body reports "rate_limit": "1 order/second" — that literal string is what the server emits and is kept verbatim below; the limit itself counts every request, not just order writes.

How Token Bucket Works

You start with 10 tokens (burst capacity). Each request consumes 1 token. Tokens refill at 1 per second, up to a maximum of 10.

Example Timeline

t=0s: You have 10 tokens, send 10 requests (burst). You now have 0 tokens.
t=1s: 1 token refilled. You can send 1 more request. You now have 0 tokens.
t=2s: 1 token refilled. You can send 1 more request. You now have 0 tokens.
Continue indefinitely at 1 request/second steady state.

Rate Limited Response (429)

Response Body

{
  "error": "rate limit exceeded",
  "retry_after": 0.82,
  "rate_limit": "1 order/second",
  "burst_capacity": 10
}

Response Headers

Retry-After: 1
X-RateLimit-Limit: 1
X-RateLimit-Remaining: 0
X-RateLimit-Reset: 1234567890

Header	Description
`Retry-After`	Seconds to wait before retrying (rounded up from retry_after in body)
`X-RateLimit-Limit`	Rate limit in requests/second (always 1)
`X-RateLimit-Remaining`	Available tokens right now (0 if rate limited)
`X-RateLimit-Reset`	Unix timestamp when the next token will be available

Best Practices

Queue orders serially: Submit one order at a time in a queue
Limit concurrency: No more than 1-5 concurrent requests
Respect Retry-After: Wait the full duration specified in the header before retrying
Cache tokens: Reuse your OAuth token for the full 1-hour lifetime
Monitor error rate: Alert if 429 responses exceed 5-10% of your traffic
Use exponential backoff: After hitting the limit, wait 1s, then 2s, then 4s for subsequent retries

Common Scenarios

Small Integration (<10/day)

You create a few orders per day. Rate limiting won’t affect you since you’ll process at well below 1 order/second.

Bulk Import (100+ orders)

Queue orders and process serially:

Use your 10-token burst for the first 10 orders
Then process at 1 order/second = ~90 more seconds for remaining 90 orders
Total: ~100 seconds for 100 orders

Spiky Traffic (Webhooks)

If you receive webhook bursts:

You can handle 10 concurrent orders immediately (burst capacity)
Queue additional orders and process at 1/second
Don’t try to send all at once or you’ll hit the limit

Requesting Higher Limits

Contact api-support@equosnine.com if you need higher limits for high-volume integrations. We may be able to accommodate enterprise-scale workloads.

Developer Docs