Chapter 5. Alerting on SLOs
This chapter explains how to turn your SLOs into actionable alerts on significant events. Both our first SRE book and this book talk about implementing SLOs. We believe that having good SLOs that measure the reliability of your platform, as experienced by your customers, provides the highest-quality indication for when an on-call engineer should respond. Here we give specific guidance on how to turn those SLOs into alerting rules so that you can respond to problems before you consume too much of your error budget.
Our examples present a series of increasingly complex implementations for alerting metrics and logic; we discuss the utility and shortcomings of each. While our examples use a simple request-driven service and Prometheus syntax, you can apply this approach in any alerting framework.
In order to generate alerts from service level indicators (SLIs) and an error budget, you need a way to combine these two elements into a specific rule. Your goal is to be notified for a significant event: an event that consumes a large fraction of the error budget.
Consider the following attributes when evaluating an alerting strategy:
The proportion of events detected that were significant. Precision is 100% if every alert corresponds to a significant event. Note that alerting can become particularly sensitive to nonsignificant ...