Book
Observability Engineering
Charity Majors, Liz Fong-Jones, George Miranda
Summary
Written by three engineers who built Honeycomb around this exact problem, this book argues that metrics, logs, and traces, the traditional "three pillars," are no longer enough to understand systems built from dozens of interacting services. It replaces that model with high-cardinality structured events as the core unit of observability, then works through instrumentation, the core analysis loop, and tying observability to SLO-based alerting. It reads as a practitioner's argument for a different default, not a survey of tools.
Target Readers
- SRE and platform engineers who have hit the limits of dashboard-and-alert monitoring on a microservices architecture
- Teams evaluating whether to migrate from metrics-and-logs tooling toward structured event data, and want the reasoning before the migration
- Engineers responsible for defining or tuning SLOs who need the analytical model behind them, not just the math
Tags
Colophon
- Publisher
- オライリー・ジャパン
- ISBN
- 978-4814400126
- Published
- Jan 2023
- List price
- ¥3,960incl. taxMay differ from the actual selling price on Amazon
Get this book
* The link above is an advertisement via Amazon Associates.Related Books
Prerequisites
- Recommended
Practical Monitoring
Mike Julian
Reason: After designing 'watch predefined metrics' monitoring with Julian's 'Practical Monitoring', advance to observability, which lets you explore even unknown failures. Majors et al.'s 'Observability Engineering' explains systems where high-cardinality events let you ask 'why did it happen?' after the fact, going beyond the limits of monitoring.
- Recommended
Site Reliability Engineering
Betsy Beyer, Chris Jones, Jennifer Petoff, Niall Richard Murphy
Reason: SRE presupposes 'knowing the exact state of the system' to meet SLOs, but the SRE book itself stays at the philosophy of monitoring. 'Observability Engineering' supplements the techniques—distributed tracing, high-cardinality events—to explore unknown failures, satisfying at the implementation level the observation capability SRE demands.
- Recommended
The Site Reliability Workbook
Betsy Beyer, Niall Richard Murphy, David K. Rensin, Kent Kawahara, Stephen Thorne
Reason: Once you put SLOs into operation with 'The Site Reliability Workbook', you need a foundation to measure SLIs accurately and trace the causes of violations. 'Observability Engineering' provides that measurement-and-investigation foundation, concretizing the observation infrastructure that supports SLO-based operations.
- Recommended
Release It!, 2nd Edition
Design and Deploy Production-Ready Software
Michael T. Nygard
Reason: After implementing stability patterns like circuit breakers from Nygard's 'Release It!', you must observe whether they actually work in production. 'Observability Engineering' makes the activation of those patterns and the system's internal behavior visible, making it verifiable that the 'unbreakable design' is functioning.
- Recommended
Cloud Native DevOps with Kubernetes
John Arundel, Justin Domingus
Reason: Operating many microservices on K8s, traditional monitoring can no longer trace 'where and what happened'. 'Observability Engineering' makes the internal state of distributed systems visible via distributed tracing and structured events, providing the observation capability essential to cloud-native operation.
- Related
Building Secure and Reliable Systems
Best Practices for Designing, Implementing, and Maintaining Systems
Heather Adkins, Betsy Beyer, Paul Blankinship, Piotr Lewandowski, Ana Oprea, Adam Stubblefield
Reason: Even if you design a secure and reliable system, you cannot notice breaches or degradation without continuously observing its state. 'Observability Engineering' provides the techniques to make the internals of distributed systems visible, connecting secure-and-reliable design to operation where 'whether it is actually upheld' can be detected.
Next Books
- Recommended
LLMOps
Abi Aryan
Reason: Where 'Observability Engineering' teaches observability for distributed systems in general via structured events and distributed tracing, 'LLMOps' applies that thinking to running LLM-based applications, connecting it to LLM-specific observation challenges like evaluation-metric design and early detection of hallucination.
Sources