Chapter 1. The Shift to Agentic Engineering
For most of the last decade, site reliability engineering (SRE) has been framed as a scaling problem with a known solution. Automate more, alert better, document harder, and hire smarter engineers. That framing worked, until it stopped working. The problem today is not that systems are unreliable. It is that they have grown beyond what humans can reliably reason about in real time.
Consider a senior on-call engineer at 03:00. A latency alert fires on a checkout path. The runbook calls for a five-step diagnostic: recent deployments, dependency health, traffic pattern, regional capacity, and configuration drift. The engineer is good at it, which is exactly why they are on the rotation. They pivot through four dashboards to find recent changes, cross-reference a deployment with a config push, scan a Slack thread for a rumor about a noisy neighbor in another region, and run a topology lookup that takes 90 seconds because the metadata service is in the same region that just degraded. By the time they have a working ...
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