Observability Engineering, 2nd Edition

Part II. Instrumentation Fundamentals

Part I covered the concepts you need to understand observability. This part is where you start building it.

Instrumentation is the process of adding code to your systems so that they emit telemetry—and getting that right is the foundation that everything else rests on. This section is for most readers.

Chapter 4 introduces instrumentation and telemetry basics: what engineers should be thinking about when they instrument their code, and a first look at OpenTelemetry, the open source framework that’s become the standard for doing so.

Chapter 5 challenges the three pillars model—the idea that logs, metrics, and traces are fundamentally separate things that need to be managed separately. This chapter breaks down why that framing is limiting, and shows how all three can be unified into a single substrate: the arbitrarily wide structured event. The key insight is that predefining the shape of your telemetry predefines the questions you can ask of it. Unifying your data frees you from that constraint—capture all the context now, and then let the investigation shape the query later. That idea is central to everything that follows. Jeremy Morrell contributed greatly to this chapter.

Chapter 6 is a guest chapter written entirely by Jeremy Morrell. It walks through common system conditions to show concretely what telemetry you should be capturing and what questions you can answer as a result. This is practical, hands-on guidance for how to start instrumenting ...