"Make it reliable" is not a goal you can measure. SRE replaces it with numbers. A service level indicator (SLI) is a measurement of how the service is behaving, such as the share of requests that succeed or the time they take. A service level objective (SLO) is the target you set for that measurement, and a service level agreement (SLA) is the promise made to customers, usually with consequences attached.
The idea that ties them together is the error budget: the amount of unreliability an SLO allows. While budget remains, teams can ship features and take risks. When it runs out, the focus shifts to stability. This gives developers and operators a shared, data-driven way to decide how fast to move, instead of arguing about it.
With targets and budgets defined, you'll be ready for the next module, Incident Management & Postmortems, which covers how to respond when a service falls short and how to learn from it afterwards.
A side-by-side look at the four terms at the heart of SRE measurement, and how each one relates to the others.
How an error budget helps teams decide when to ship new changes and when to slow down and focus on reliability.
A conference talk from a Google engineer on using an error budget to move fast without giving up high reliability.