Website Care, Performance & Security
Service Level Indicator (SLI)
A service level indicator (SLI) is a clearly defined numerical measure of a service behavior that matters to users, such as successful requests or response latency.
An SLI makes a service's reliability observable rather than leaving it at a vague promise such as “the website is fast.” For a booking system, the team might define the indicator as the proportion of valid booking requests that return a successful response within two seconds. A complete definition identifies the event being counted, the denominator, exclusions, where measurements are collected and the observation period. These choices matter: monitoring only the server can miss slow experiences on mobile devices.
An SLI is a measurement, not a target. A service level objective (SLO) states the acceptable result, for example that 99.5% of eligible requests should complete within the specified time across 30 days. A service level agreement (SLA) can set contractual consequences, but is not required for internal reliability work. Teams often track request success and latency and, for data services, durability or freshness. An error budget can then translate SLO shortfalls into operational decisions.
Avoid measuring every easy-to-count event. Choose a small number of indicators that map to user outcomes, specify how missing telemetry is treated, and review the definition after major architectural changes. A dashboard showing 99.9% uptime is not meaningful without knowing what counted as available.