TL;DR
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
The Pragmatic Engineer published excerpts from Sam Newman’s 2026 book, Building Resilient Distributed Systems, on disaster recovery and designing resilient AI-powered software. The material discusses recovery targets, failover patterns, infrastructure constraints and uncertainty around AI models and vendors; it does not provide evidence that outages have increased because of AI coding tools.
The Pragmatic Engineer has published excerpts from software engineer Sam Newman’s 2026 book, Building Resilient Distributed Systems, covering disaster recovery and the challenges of building dependable AI-powered software. The material sets out recovery objectives and failover options while arguing that teams must plan for infrastructure failures, changing models and uncertainty about vendors.
In the excerpt from Chapter 10, Newman describes disaster recovery (DR) as the procedures used when a catastrophic event disrupts a substantial part of a system. Examples include fire, flooding, power problems and failures in cloud infrastructure. He says recovery plans require trade-offs: reducing acceptable downtime or potential data loss generally increases cost and complexity.
Newman recommends defining two measurable targets. A recovery point objective (RPO) sets the maximum acceptable period of data loss, based on how recently information was backed up or synchronized. A recovery time objective (RTO) sets the acceptable time for restoring service after an incident, including restoring data, making computing resources available and redirecting traffic.
The newsletter says the book also covers backup-and-restore, pilot-light, warm-failover and active-active recovery patterns, along with disaster recovery objectives and service-level objectives. Its AI-related topics include model-provider resilience, changes to models, observability, multi-model and multi-vendor strategies, transparent costs and regular vendor reviews. The supplied excerpt does not specify detailed implementation procedures for those AI practices.
Recovery Targets Shape Resilience
The guidance matters because a system’s reliability depends not only on preventing individual service failures but also on how quickly it can recover when a larger outage overwhelms those safeguards. RPO and RTO turn broad expectations into targets that engineering and business teams can use to compare recovery plans, although a target alone does not prove that a system can meet it.
For AI-powered products, the newsletter highlights additional dependencies: an application may rely on external model providers, and the models themselves can change. Newman’s outlined emphasis on observability, multiple providers and cost visibility points to planning for disruptions or changes beyond a software team’s direct control. The excerpt presents these as design considerations, not as guarantees that a particular architecture will prevent outages.
As an affiliate, we earn on qualifying purchases.
From Site Outages to AI Dependencies
The material appeared in The Pragmatic Engineer after a podcast conversation with Newman, who had recently published the book through O’Reilly Media. The newsletter says it received permission to share excerpts from two chapters, and identifies the supplied disaster-recovery passage as coming from Chapter 10. The book has 16 chapters, according to the report.
The report also says Newman spent three years writing the book, rather than the two he initially planned, and revisited sections to account for AI. The newsletter’s introduction links AI coding agents to more code deployments and production bugs, but this is the author’s framing; the source provides no outage data or independent evidence establishing that AI tools have caused an increase in incidents.
“As you reduce your period of acceptable downtime or data loss, cost and complexity increase.”
— Sam Newman, in the excerpt from Building Resilient Distributed Systems
As an affiliate, we earn on qualifying purchases.
Outage Evidence and Recovery Results
The source does not establish whether AI coding agents have caused more production outages, nor does it provide statistics on incident rates, deployments or their causes. That connection appears in the newsletter’s introduction and should be treated as its framing, not as a demonstrated finding. The supplied passage also does not quantify GPU or CPU shortages or identify particular affected providers.
It remains unclear what specific RPO or RTO targets Newman recommends for different kinds of systems, and the provided excerpt ends before a complete discussion of recovery approaches. The report does not compare the cost or effectiveness of individual patterns, or show that any named organization has adopted them successfully. Its descriptions are guidance from a book excerpt, not reported test results.
distributed system resilience tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Applying the Book’s Recovery Guidance
The immediate next step for readers is to use the book’s framework to define acceptable data loss and downtime for their own systems, then evaluate recovery options against those requirements and their cost and complexity. Newman’s chapter covers a range of patterns, but teams would need to test their own plans to learn whether recovery targets are achievable.
The newsletter says the book is available from its publisher and recommends it to engineers responsible for reliable systems. The source gives no scheduled follow-up announcement, deployment or industry-wide milestone. Further details on the AI-resilience topics would depend on the full chapters or subsequent reporting.
cloud infrastructure backup and restore
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
What is the news?
The Pragmatic Engineer published authorized excerpts from Sam Newman’s 2026 book on disaster recovery and resilient AI-powered software.
What do RPO and RTO mean?
RPO is the maximum acceptable period of data loss, while RTO is the acceptable time to restore service after an incident.
Which recovery approaches does the report mention?
It lists backup-and-restore, pilot-light failover, warm failover and multisite active-active as patterns covered by the book.
Does the source prove AI coding tools are causing more outages?
No. The newsletter’s introduction makes that connection, but the supplied material includes no incident data or independent evidence establishing an increase caused by AI coding tools.
What AI-specific resilience issues does the book address?
The newsletter says the book discusses model-provider resilience, model changes, observability, multi-model and multi-vendor strategies, cost transparency and vendor reviews.
Source: rss
Fall Picks
fall essentials
As an affiliate, we earn on qualifying purchases.
