Rate limits

Request limits for each Adzen API, the headers that report them, and what to retry.

Limits

Product	Limit
Ads API	600 requests per minute per endpoint, per key, and 1,200 requests per minute per key across all endpoints. Each advertiser key has its own limits. Hourly insights (`granularity=hour`) at ad set or ad level, or with the `ad_set` or `ad` breakdown, allow 30 requests per minute per key.
Publisher insights	60 requests per minute per account.
AI API	`POST /v1/ai/process` returns `429` with `Retry-After: 1` when it is busy.
UGC API	Request each post once per window. After a match, request it again after `ttl_seconds`; after a no-match response, wait `no_ad_retry_after_seconds` seconds. See No match.

Ads API

An endpoint is the method and the route, with ids as placeholders. Pausing two campaigns counts twice against POST /v1/ad-manager/campaigns/{id}/pause.
Windows are fixed clock minutes. Every count starts over at the top of each minute.
An idempotent replay counts against the limits.
Over a limit, the API returns 429 with code rate_limit_exceeded:


{
  "error": {
    "type": "rate_limit_error",
    "code": "rate_limit_exceeded",
    "message": "more than 600 requests a minute on this endpoint; retry after 12 seconds",
    "param": null
  }
}

Publisher insights

The limit counts every request for your account, whichever key sends it. The 429 uses the Publisher insights error body with type rate_limit_error and code rate_limited.

AI API

The 429 body is {"error": "endpoint overloaded, retry shortly"}. Wait one second, then send the request again. AI API responses carry no X-RateLimit-* headers.

Headers

Header	Ads API	Publisher insights
`X-RateLimit-Limit`	Requests allowed in the current minute	Requests allowed in the current minute
`X-RateLimit-Remaining`	Requests left in the current minute	Requests left in the current minute
`X-RateLimit-Reset`	Unix time, in seconds, when the minute ends	Seconds until the minute ends
`Retry-After`	On `429`: seconds until the next minute, at least 1	On `429`: seconds until the minute ends

The Ads API reports whichever of the per-endpoint and per-key limits has fewer requests left. The headers do not reflect the 30-requests-per-minute hourly insights limit. On the Ads API, every response to an authenticated request carries the X-RateLimit-* headers. On Publisher insights, successful responses and 429 responses carry them.

Retries

Response	What to do
`429`	Wait `Retry-After` seconds, then send the same request again.
`409` `idempotency_key_in_use` (Ads API)	The first request with this `Idempotency-Key` is still running. Wait `Retry-After` seconds, then retry with the same key.
`503` with `Retry-After`	Wait `Retry-After` seconds, then retry.
Other `5xx`	Retry with exponential backoff and jitter. Cap the number of attempts.
Other `4xx`	Do not retry unchanged. Fix the request first.

On the Ads API, send the same Idempotency-Key on every retry of a POST. See Idempotency.