Monitoring the performance of APIs is a common part of operating internal and external APIs. Select operations within an API can have simple latency tests, as well as more robust load and network testing that will define the overall performance of APIs. Performance testing should be present for all APIs, with the bounded context and frequency of performance testing vary from API to API.
Their are numerous considerations regarding the performance of an API, including network and regional considerations, as well as backend, dependencies, and other more complex considerations. Simple recurring performance tests that record latency and response time is where performance monitoring should begin, then expanded upon once performance testing reaches near 100% coverage across APIs.
As performance testing is setup and regularly reviewed, think about the business implications of performance and how performance monitoring, results, and outcomes impact the bottom line. Do the work to separate performance and real-time needs, as well as the overall complexity and scope of API design and data volume. Performance impacts almost every stage of the API lifecycle, but is often just associated with API runtime once it begins to impact business.