Part IV. From SRE to ARE
Three chapters plus the Epilogue. The operating model, the multiagent architecture, the ecosystem of named agents, and a closing reflection on the human signal.
Parts I through III built the architectural substrate. The discipline of agentic reliability engineering, the four planes, the maturity model, the five flagship SLOs, the four workflows of Part II, the vocabulary and patterns of Part III, and the governance and simulation that close the substrate’s loop. Approximately 350 pages of architecture. What that architecture cannot do is run itself. Production reliability has never been just a property of the system; it has always been a property of the system plus the team that operates it, and the agentic era does not change this. What changes is what the team is, what the team does, and how the team’s work composes with the architecture’s work. Part IV contains the chapters where that human side gets its formal treatment.
The premise that ties the three chapters together is that the technical shift requires an organizational shift. A team that adopts the architecture of Parts I, II, and III but keeps its operating model unchanged finds that the architecture’s value is bounded by what the old operating model can absorb. The SRE who used to be paged at 4 a.m. to roll back a deploy is now reading the architecture’s Why Trace at 11 a.m. to understand why a rollback happened; the on-call rotation that used to be measured in minutes to acknowledge is now ...
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