The Site Reliability Workbook

Chapter 14. Configuration Design and Best Practices

By Štěpán Davidovič with Niall Richard Murphy, Christophe Kalt, and Betsy Beyer

Configuring systems is a common SRE task everywhere. It can be a tiring and frustratingly detailed activity, particularly if the engineer isn’t deeply familiar with the system they’re configuring or the configuration hasn’t been designed with clarity and usability in mind. Most commonly, you perform configuration in one of two scenarios: during initial setup when you have plenty of time, or during an emergency reconfiguration when you need to handle an incident.

This chapter examines configuration from the perspective of someone who designs and maintains an infrastructure system. It describes our experiences and strategies for designing configuration in a safe and sustainable way.

What Is Configuration?

When we deploy software systems, we do not think of them as fixed and never-changing. Ever-evolving business needs, infrastructure requirements, and other factors mean that systems are constantly in flux. When we need to change system behavior quickly, and the change process requires an expensive, lengthy rebuild and redeployment process, a code change won’t suffice. Instead, configuration—which we can loosely define as a human-computer interface for modifying system behavior—provides a low-overhead way to change system functionality. SREs take advantage of this on a regular basis, when deploying systems and tuning their performance, as well as during ...