Chapter 3. The Agentic Resilience Engine
Self-healing systems are not systems that recover. They are systems that learn.
That sentence is the chapter’s whole argument, and the rest of the chapter is the engineering work it implies. Reliability has long been framed as prevention. Design for redundancy. Harden the architecture. Eliminate single points of failure. Reduce mean time to recovery. These practices have served engineering teams well, and they remain useful, but they sit on an assumption that is increasingly hard to defend. They assume that the failure modes a system will encounter can be enumerated ahead of time. Modern systems are too complex, too interconnected, and too dynamic for that assumption to hold. Dependencies shift at runtime. Load profiles change without warning. Data quality fluctuates. Models drift. Under these conditions, resilience based purely on anticipation becomes fragile.
The shift this chapter argues for is not a small one. It moves resilience from a structural property to a living capability. A traditional resilient system ...
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