Part 1 of 8 · 1 min read
Why this page comes second
In this module you will
- Say what makes an alert useful, and why most alerts should be built on an SLO.
- Explain precision and recall for alerts, and the trade-off between them.
- Compute a burn rate and the alert threshold for a chosen share of the error budget.
- Design three kinds of dashboard for three audiences, and lay out a chart well.
- Describe a healthy on-call rotation and what to do first when an alert fires.
Module 1 produced numbers that say whether users are well served. This page turns those numbers into two tools: alerts that call a person, and dashboards that show health at a glance. Both come from "Understanding Distributed Systems" (UDS), chapter 20. Numbers marked "made up" only show the idea.