SLO
Definition
Service Level Objective - specific measurable target for service performance, helping organizations maintain quality and reliability in services.
Use Cases
- Google: Setting reliability targets for user-facing services to balance feature velocity with reliability. — Google popularized SRE practices that use SLOs tied to SLIs (like request latency and error rate) and manage an error budget. Teams monitor SLIs, compare them to SLO targets, and use the remaining error budget to decide whether to prioritize reliability work or ship new features. (More consistent service reliability and clearer decision-making tradeoffs between launching changes and maintaining stability (via error-budget-based governance).)
- Netflix: Maintaining streaming and API reliability at global scale while deploying changes frequently. — Netflix uses extensive observability (metrics, logs, tracing) and automated operations practices (including progressive delivery and resilience testing). SLO-style targets for key user journeys and APIs are monitored to detect regressions and guide operational focus. (Improved ability to detect and respond to performance/reliability issues quickly, supporting frequent deployments while protecting customer experience.)
Frequently Asked Questions
- What's the difference between an SLO and an SLA?
- An SLO is an internal (or team-level) measurable reliability goal, like "99.9% of requests succeed" or "p95 latency under 200 ms." An SLA (Service Level Agreement) is a customer-facing contract that may include penalties or credits if the provider fails to meet it. In practice, SLAs are often looser than internal SLOs so teams have a safety margin.
- When should I use SLOs?
- Use SLOs when you run a service where reliability matters and you need a clear target to guide engineering and operations. They are especially useful when you have multiple teams, frequent deployments, or recurring incidents, because SLOs turn reliability into a measurable goal and help prioritize work (for example, pausing risky releases when the error budget is low).
- How much do SLOs cost?
- Defining an SLO is free, but measuring and reporting it costs money through observability tooling. Costs depend on what you measure and how: metrics ingestion and retention, log volume, trace sampling, dashboarding, alerting, and any managed SLO features in your monitoring platform. Higher request volume, higher-cardinality labels, longer retention, and more detailed tracing generally increase cost.
Category: software
Difficulty: intermediate
Related Terms
See Also