A service-level-indicator is a specific quantitative measure used to track the performance of a service along a single dimension — such as request latency, throughput, error rate, or availability. It is the raw metric that is continuously collected by monitoring infrastructure, visualized in dashboards, and compared against a service-level objective to determine whether the service is meeting its reliability target. An SLI is defined by three parameters: the metric type (what is measured), the measurement method (how it is sampled and computed), and the reporting interval. Its persistence mechanism is continuous data collection through monitoring agents and aggregation pipelines, stored in time-series databases.
[formal: sli | substrate: mind | horizon: hours | explicit: yes | epoch: 0.01]