Observability Engineering, 2nd Edition

Chapter 1. What Is Observability?

The topic of observability has generated a steady supply of attention and buzz over the past 10 years, ever since it was first introduced to the field of software development. However, as with most complex topics, a surging tide of practitioner adoption and interest in observability was swiftly countered by a wave of oversimplification and appropriation, fueled by hype and marketing budgets.

When you see this cycle play out repeatedly, it is easy to be cynical: maybe words just don’t mean things when money gets involved. It’s worth taking a step back to consider: why did we all get so excited in the first place? What was so revelatory, and why did it matter?

This chapter examines the mathematical origins of the term “observability,” and how the software development industry has interpreted it and adapted it to our world. We also look at the tectonic plates that are colliding at almost every company right now, between the old ways (which evolved out of server ops and still dominate the ecosystem) and the new ways—derived from first principles and far more expansive in capabilities, yet swimming upstream against the way things have always been done.

Along the way, we’ll argue that observability isn’t just a debugging technique or a category of tooling; it’s a property of software dependability, as fundamental as reliability or availability.

Observability as a concept was born out of necessity, when traditional tools and techniques simply could ...