SLI: service level indicator
What you measure, such as the percentage of requests that succeed, or that respond within 300 ms.SLO: service level objective
Your internal target for that measure, for example 99.9% of requests succeed over 30 days.SLA: service level agreement
The promise to customers, usually written into a contract, with penalties such as refunds if it's missed.Error budget
The failure you can afford. With a 99.9% SLO you can fail 0.1% of the time, about 43 minutes a month. Spend it on shipping features; when it runs out, focus on reliability.Golden rule
Promise less than you aim for: keep the SLA below the SLO.
Example: a payments API
- SLI: percentage of payment requests that succeed
- SLO: 99.9% over 30 days
- SLA: 99.5% promised to customers
- Error budget: about 43 minutes a month
Why it matters for your career
SLOs and error budgets are core to SRE roles, and increasingly to DevOps and platform engineering too.
