...
Implications on back-end services/systems
Maximum
...
backend throughput: API Publisher
According to the API Gateway architecture, the APIs in the Gateway are actually proxies to an actual service hosted within your organization, cloud, etc. This usually means that there is a physical capacity that your back-end backend services can handle. Although you expose your API on defined limits (subscription tiers), as the number of applications that consume your API grows, the number of requests being served by your API rise, which in turn means that the number of requests served by your back-end backend system rise as well. Therefore, although none of the applications may exceed their own allocated quotas, their combined load might hit the maximum capacity that can be handled by your back-end backend system. To prevent your back-end backend system from getting overloaded, the limits enforced by the Maximum Backend Throughput act as a hard stop on the number of requests that your back-end backend system can serve within a given time period. The counters maintained when evaluating the maximum back-end backend throughput are shared across all nodes of the Gateway cluster and apply across all users using any application that accesses that particular API. For information on how to specify maximum backend throughput limits, see Setting Maximum Backend Throughput Limits.
Implications on the APIs in the Gateway
...