Rate Limits & Retries

Rate limits

Augment APIs allow up to 10 requests per second (600 requests per minute) per customer, unless a different limit is explicitly documented for a specific endpoint.

Retrying failed requests

Only retry when the failure is likely transient. A good rule of thumb: if the same request would probably succeed without any changes, it's worth retrying.

Retry these responses

Status / Condition	Reason
500 Internal Server Error	Unexpected server-side failure
502 Bad Gateway	Upstream service unavailable
503 Service Unavailable	Server temporarily unable to handle the request
504 Gateway Timeout	Upstream service did not respond in time
Network-level failures	Timeouts, connection resets, refused connections

Do not retry these responses

Status	Reason
400 Bad Request	Request is malformed or missing required fields
401 Unauthorized	API key is invalid, missing, or revoked
403 Forbidden	Authenticated but not authorized for this action
404 Not Found	Resource does not exist
422 Unprocessable Entity	Request is well-formed but fails validation

These errors indicate a problem with the request itself. Retrying without changing the request will produce the same result.

We monitor 4xx and 5xx errors across all API traffic. If you're running into persistent errors or need help troubleshooting, don't hesitate to reach out to your customer success representative or our integrations team.

Backoff strategy

Use exponential backoff with full jitter. Suggested parameters:

Parameter	Value
Initial base delay	~500 ms
Maximum delay cap	~30 seconds
Maximum attempts	5
Maximum total wait	~2 minutes

Stop retrying when you hit either the attempt limit or the total time budget — whichever comes first.