Similarly to the ELB and CloudFront, we can monitor the performance of the API gateway by selecting the metrics related to each API. In this example, we have chosen to select all the metrics of the Test API. The API was tested externally during a 15-minute EST window and as we can see, traffic was recorded during that time. We have chosen to examine both 400 and 500 error codes, the latency of the API, and the integration latency, which measures the performance from the API gateway to the backend. We also have a count metric selected, but the counts are really small so they are not registering well on this chart:
However, what we can see from this chart is a nice correlation between the integration latency and the overall latency of the API. This means that our API is performing well and the larger chunk of the latency is due to the backend responding...
