TL;DR
Get business pricing on tech for your team
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
The Pragmatic Engineer published excerpts from Sam Newman’s 2026 book, Building Resilient Distributed Systems, covering disaster recovery and resilience for AI-powered software. The material outlines recovery targets and infrastructure trade-offs; it does not provide evidence that AI coding agents have caused a measured rise in outages.
The Pragmatic Engineer has published excerpts from Sam Newman’s 2026 book, Building Resilient Distributed Systems, addressing how software teams can plan for major outages and manage resilience risks in AI-powered systems. The material covers disaster recovery targets, recovery patterns, infrastructure availability and the challenges of relying on changing AI models and providers.
Newman’s excerpt on disaster recovery focuses on failures that can take out a whole site, such as a cloud region or data center, rather than a single service instance. Fires, floods, electrical problems and failures in cloud infrastructure can overwhelm protections designed for smaller disruptions. Disaster recovery, the book says, requires procedures for restoring services and data after such events.
The excerpt identifies two measures for setting recovery requirements. A recovery point objective (RPO) sets the maximum acceptable period of potential data loss, based on when data was last backed up or synchronized. A recovery time objective (RTO) sets how long a service can remain unavailable. Meeting an RTO can involve restoring data, making computing resources available and redirecting traffic.
The Pragmatic Engineer’s overview says the book also discusses backup-and-restore, pilot-light, warm-failover and active-active approaches. Its AI-related topics include dependence on model providers, changes to models, observability, and ways of managing uncertainty, such as using multiple models or vendors, examining costs and regularly reviewing vendors. Newman’s book took three years to complete, rather than the two he had originally planned, according to the publication, which says he reworked material to account for AI.
Recovery Targets Shape Resilience
The excerpt’s practical point is that resilience involves explicit trade-offs. A shorter acceptable outage or smaller tolerance for data loss can require more infrastructure, testing and operational complexity. Teams that define RPO and RTO can connect their recovery design to what their services and users can actually tolerate, rather than treating disaster recovery as a single standard solution.
The AI discussion adds dependencies that may change over time. If an application relies on an external model provider, availability and model changes can affect service behavior. The publication’s summary says Newman recommends observability and considering multiple providers or models. These are presented as design considerations in the book, not as proof that any particular architecture prevents outages.
The overview also flags constraints on access to GPUs and CPUs. A recovery plan that assumes replacement capacity will always be available may not match vendor or market constraints. Planning for which services deserve scarce resources can matter when normal infrastructure is disrupted.
As an affiliate, we earn on qualifying purchases.
From Site Failures to AI Dependencies
The report frames the book as a response to the work of building systems that remain reliable under failure. Newman’s disaster-recovery chapter extends resilience beyond individual instances to broad outages affecting an entire site. It uses “site” as a general term that can refer to a cloud region, data center or comparable infrastructure location.
The Pragmatic Engineer says the book spans 16 chapters and that its published excerpts cover only part of the disaster-recovery material and selected AI-relevant themes. The source also says Newman discussed the book on the publication’s podcast the previous week. Its editor disclosed a personal recommendation of the book and said the recommendation was not paid and its links were not affiliate links.
“The recovery point objective (RPO) defines the maximum amount of time that can have passed since the data was last backed up or synchronized.”
— Sam Newman, in the book excerpt published by The Pragmatic Engineer
cloud infrastructure monitoring tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Outage Trends Remain Unverified
The report says software teams are seeing more outages and links that concern to increased code deployments involving AI agents and more production bugs. It provides no outage data, time period or comparison baseline to substantiate that trend, so it should be treated as the publication’s framing rather than a measured finding. The excerpt also does not quantify the effect of AI tools on incident rates.
The supplied material does not provide detailed evaluations of particular disaster-recovery patterns, vendors or model providers. Nor does it specify which AI services or infrastructure shortages are most affected. The appropriate recovery targets and architecture will depend on each system’s needs and constraints.
As an affiliate, we earn on qualifying purchases.
Applying the Guidance to Systems
The immediate next step for readers is to use the book’s framework to define acceptable data loss and downtime, then compare those requirements with the cost and complexity of available recovery patterns. Teams can also examine whether their plans depend on infrastructure capacity or AI providers that may not be available during a disruption.
The published material is an excerpt, not a report of a new industry standard or a scheduled policy change. The Pragmatic Engineer points readers to Newman’s full book for further material. The source does not give a date for additional excerpts or any forthcoming update.
As an affiliate, we earn on qualifying purchases.
Key Questions
What is the news?
The Pragmatic Engineer published excerpts from Sam Newman’s 2026 book, Building Resilient Distributed Systems, on disaster recovery and AI-related resilience.
What do RPO and RTO mean?
RPO is the maximum acceptable period of potential data loss. RTO is the maximum acceptable time a system can remain unavailable after an incident.
What AI resilience issues does the book address?
The publication’s summary lists model-provider resilience, model changes, observability, infrastructure availability and uncertainty around vendors. It also mentions considering multiple models or providers and reviewing costs and vendors.
Does the report prove AI coding agents are causing more outages?
No. The report raises that concern but provides no incident statistics, comparison baseline or time window to establish that AI agents have increased outage rates.
What is not yet known?
The excerpt does not quantify how often large-scale failures occur, how AI affects incident rates, or how much particular recovery approaches cost. Those factors depend on the system and its providers.
Source: rss
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
