Part 1 of 8 · 1 min read
Why this page comes first
In this module you will
- Define a service level indicator (SLI) as a ratio of good events to total events.
- Pick an SLI that reflects what users feel, and decide where to measure it.
- Set a service level objective (SLO) and work out its error budget.
- Explain why averages mislead and why a small slow fraction can double the cost of a service.
- Use the error budget to decide between new features and repair work.
Level 8 built defences. This level is about running the system after launch, and every part of it needs one thing: an agreed definition of "working". Without numbers, alerts are guesses and arguments about priorities never end. The book used is "Understanding Distributed Systems" (UDS), chapter 20, with percentile ideas from "Designing Data-Intensive Applications" (DDIA). The numbers are made up to show the idea, unless the text names a book.