The Site Reliability Workbook

What the workbook is

The Site Reliability Workbook is the second of Google's SRE books, published by O'Reilly in 2018. The first book, Site Reliability Engineering, explains the principles. The workbook exists because readers kept asking the obvious follow-up: how do I actually do this on my team? Every chapter answers that question with worked examples, configuration snippets and case studies, several of them from companies other than Google. Google publishes the full text free at sre.google under a Creative Commons licence (CC BY-NC-ND 4.0).

What the chapters cover

• Foundations: how SRE relates to DevOps, implementing SLOs, SLO case studies, monitoring, alerting on SLOs, eliminating toil and simplicity.
• Practices: on-call, incident response, postmortem culture, managing load, non-abstract large system design, data processing pipelines, configuration design, and canarying releases.
• Processes: identifying and recovering from overload, the SRE engagement model, taking SRE beyond your walls, SRE team lifecycles, and organisational change management.

Is it worth reading?

Yes, if you run or plan to run anything in production. The chapters on implementing SLOs and alerting on SLOs are the ones people bookmark: they walk from "what should we measure" to an error budget policy and then to burn rate alerts, trying several alerting approaches and explaining why each one falls short before landing on multiwindow, multi burn rate alerts. That progression teaches you how to reason about alerting, not just which rule to copy. The on-call, incident response and postmortem chapters are equally concrete, with example documents you can adapt for your own team.

Prerequisites and time

You should be comfortable with how a web service is deployed and monitored: requests, latency, errors, dashboards. You do not need to have read the first SRE book, although chapters refer back to it and it helps to skim its principles section. Reading the whole workbook carefully takes a few weeks at a chapter or two a sitting. If time is short, read the SLO, alerting, on-call, incident response and postmortem chapters first; together they are the core of the job.

Free or paid?

The online edition is free and complete. O'Reilly sells the printed and ebook editions if you prefer those.

What you can do after it

Write SLOs for a real service from what users experience, set an error budget policy your team will actually follow, replace noisy threshold alerts with burn rate alerts, run an incident with clear roles, and write a blameless postmortem that leads to fixes rather than blame.

Common questions

Is it only relevant at Google scale? No. The SLO and alerting chapters apply to a single service with a handful of users, and the case studies deliberately come from smaller organisations too.

Is it outdated because it is from 2018? The tooling examples age, but the reasoning about SLOs, toil, on-call load and postmortems has not changed, and it is still the reference most SRE teams quote.

Best for: backend, DevOps and platform engineers who want to move from keeping things up by instinct to running services against explicit reliability targets.