That link is documenting the shape of a specific request ("wait for operation") sent to the Service Management API. This is tangentially related to API Gateway, but is not at all related to the timeout settings of your gateways. As Todd said, you are right to be confused on how to configure this because it is not yet configurable. This is on the roadmap, and as mentioned we also have a FR to track your specific request internally.
In the meantime, we can actually raise the default timeout (within reason) without much work. Would this help in the short-term?