Site Reliability Engineering, 2nd Edition

Chapter 8. Service Level Objectives

By Alex Bramley, Brett Beekley, Chris Jones, David Huska, Derek Remund, Ethan Truong, Faraz Khan, Jose Andrade, Micah Lerner, Niall Richard Murphy, Parker Roth, and Yuri Grinshteyn with Carmela Quinito

The first edition of Site Reliability Engineering did much to popularize the SLO framework and kick-start its adoption across the wider computer industry. Others promptly took the baton and ran with it, producing a wide array of introductory, intermediate, and advanced material in almost every form imaginable. This reflects the continuing need for effective models at all levels of experience to help people understand the behavior of the systems they’re responsible for, the level of service they should provide to users, how much effort is justified to achieve it, and what to do when those expectations aren’t met. As we were writing this edition, SLOs remained the most frequently queried topic in the online version of the first edition.

The challenge for this chapter is that the market for introductions is already well served, but much of the more advanced material does not make sense without a basic understanding of SLO concepts. Therefore, we’ll start with a brief overview before turning our attention to the following topics, which were either not covered or only briefly addressed the first time around:

• Pitfalls in SLO adoption

• Critical user journeys

• Subordinate SLOs

• SLOs for B2B businesses (and per-customer SLOs)

• Implicit SLOs

• Client-side ...