Preface
For the past decade, we have scaled reliability by automating execution and have kept meeting systems that changed faster than we could reason about them in real time. While writing this book over the past year, I struggled to find the perfect name for the concept I am describing. I have heard it called autonomous reliability, agentic operations, self-driving infrastructure, and half a dozen other names from teams who see the same pattern in their own work. I call it agentic reliability engineering (ARE), and I believe it is what comes after site reliability engineering (SRE). These next few pages will show you why.
A bit of history first. SRE, as a named discipline, is about 20 years old. The team at Google formalized it. The Site Reliability Engineering book (O’Reilly, 2016) and The Site Reliability Workbook (O’Reilly, 2018) carried this concept into the rest of the industry. The core idea was simple: reliability is an engineering problem, not an operational afterthought. The discipline gave us error budgets, SLOs, toil reduction, blameless postmortems, and the rest of the vocabulary every modern infrastructure team now uses. It worked. It worked extraordinarily well. The systems we run today are larger, faster, and more reliable than anything that came before, because we treated their operation as engineering rather than as something to put up with.
Why I Wrote This Book
Here is what I have seen happen over the last few years in the teams I work with directly and across ...
Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.
Read now
Unlock full access