SLA, SLO, SLI — Explained with Examples

In this tutorial, you'll learn about SLA, SLO, SLI. We cover key concepts, practical examples, and best practices to help you understand and apply this topic effectively.

SLA (Service Level Agreement) is a formal contract between a provider and a customer that specifies the expected level of service, usually including penalties for breaches. SLO (Service Level Objective) is an internal target for a reliability metric. SLI (Service Level Indicator) is the actual measured metric — the raw data point.

SLIs are concrete measurements: request latency, error rate, throughput, uptime percentage. SLOs define thresholds for SLIs (e.g., "99.9% of requests complete under 200ms"). SLAs are the legal wrapper around SLOs, often with financial terms. A typical SaaS SLA might promise 99.9% uptime; if the provider falls below, customers get service credits.

Real-world analogy. SLI is your car's speedometer reading 63 mph. SLO is your personal rule: "I will stay under 70 mph." SLA is the speed limit law: the government mandates ≤65 mph on this highway, and if you exceed it, you pay a fine. You set your SLO (70) below the SLA (65? no — you'd set it lower to be safe).

Example calculation:

SLI (uptime) = (total minutes - downtime) / total minutes * 100
SLO = 99.9% uptime per month
SLA = 99.9% uptime, credits if below

Related terms: Observability, Chaos Engineering, Zero Downtime Deployment, CI/CD, Canary Deployment

Related tutorial: SLO Management Guide