Rate limits

Requests are rate limited per API key. When a limit is exceeded the API returns 429; back off and retry.

Limits#

  • Each API key carries its own per-second limit (set when the key is created or edited in Settings → API keys). There is no shared per-tenant limit.
  • The gateway also applies a product subscription (the Ocp-Apim-Subscription-Key) — see Authentication.

Handling 429#

Tip

Retry with exponential backoff and jitter. If a single key is consistently throttled, raise its limit or spread load across keys — see Errors.
Was this page helpful?