API Traffic Management 101

Chapter 3. Monitoring Traffic

In this chapter, we drill down into the basics of traffic monitoring. This means understanding the typical metrics we can monitor in order to understand that overall health of our system. We also spend some time reviewing some common traffic formulas that help you better understand the runtime condition of your service ecosystem. With these elements in hand, we can custom-build suites of monitoring metrics that will assist you in observing and monitoring the API traffic that’s important to your business.

Rather than just focus on low-level metrics like CPUs and memory, it is also important to monitor service-level metrics (e.g., number of queries processed per second) and even business-level metrics (e.g., number of new customers added this week). Being able to separate monitoring into levels helps us better focus on getting the right metrics into the right dashboard at the right time.

Monitoring Levels

An important aspect of monitoring is to ensure that you know what level you are observing and what kinds of information you want to collect at that particular level. A useful way to think about levels is to carve up your observation space into three levels: infrastructure, service, and business, as illustrated in Figure 3-1.

Figure 3-1. Three traffic monitoring levels

Infrastructure Metrics

At the infrastructure level, the focus is on the health ...