Project adding a public API. Initially only used by close partners. Once we open it up, we plan to introduce a proper API gateway to handle rate limiting, billing etc.

API will include long running AI model calls. In order to handle this and also prevent model server overload, we will need to add queuing.

Is the job of queuing long running web requests commonly part of the offering of API gateways?

Asking since we only need the API gateway in phase 2 but need queuing in phase 1.