Observability Engineering, 2nd Edition

Preface

When we wrote the first edition of this book, the observability landscape was shifting, and practices were advancing faster than we could collectively write them down. In retrospect, we were chasing a moving target. It took three and a half years to write and edit the book that was released in 2022.

In the four years since, the landscape and practices for observability have solidified a great deal, which makes it easier to write about. However, life is never boring; ChatGPT and GitHub’s Copilot were also released in 2022, kicking off an explosive period of innovation that has yet to subside.

As we write this preface, in March of 2026, the entire software industry is going through yet another unprecedented upheaval, a convulsive period of compressed change. New AI development workflows are being hammered out in real time, for durable code as well as disposable code. Harness engineering, agentic orchestration, Phoenix architectures…every day seems to come with new terms and new concepts. It feels overwhelming trying to keep up, and the only consolation is that everyone else feels just as overwhelmed.

We’ve seen AI models leap generations, entirely new categories of tooling appear, and workflows that felt cutting-edge become table stakes, all within one fiscal quarter. This is what exponential change feels like from the inside: not a smooth trajectory, but a series of jarring “wait—already?” moments. The hype trains are roaring, and it can be difficult to pick trustworthy ...