Observability Engineering book, 2nd Edition [pdf]
The second edition of "Observability Engineering," a book published by O'Reilly Media and available as a free PDF from Honeycomb, covers modern observability practices for understanding complex software systems through high-cardinality, high-dimensionality data and service-level objectives.
Background
- Honeycomb is a company that sells observability software (used by engineers to understand how their systems behave in production). This page offers a free PDF of the 2nd edition of *Observability Engineering*, a book co-authored by Honeycomb's founders.
- "Observability" (often shortened to "o11y") is a property of software systems: how well you can understand what's happening inside them from the data they emit, without having to pre-guess every possible failure mode. It goes beyond traditional monitoring (collecting fixed metrics like CPU usage) by emphasizing high-cardinality events and structured logs.
- The book popularized the "Three Pillars of Observability" (logs, metrics, and traces) in its first edition, though the 2nd edition reframes the concept around "high-cardinality events" and "generative" debugging workflows.
- This matters because as software systems grow more distributed and complex (microservices, Kubernetes, serverless), traditional monitoring tools often fail to explain novel or emergent problems. The observability approach is now an industry standard for DevOps and SRE teams.