USE is the checklist for resources: for each one, check utilization, saturation and errors.
The USE method, from Brendan Gregg, monitors every resource by three signals: Utilization (how busy it is), Saturation (how much work is queued or waiting), and Errors. It is a systematic way to find resource bottlenecks.
Utilization is the percentage of time the resource is busy. Saturation is the extra work that cannot be serviced yet and is queuing, often the earlier warning than utilization. Errors is the count of error events. Applied to every resource, CPU, memory, disk, network, USE gives a complete bottleneck checklist.
USE is resource-centric, so it shines for infrastructure: hosts, disks, network interfaces, database internals. For user-facing services, the request-centric RED method or the golden signals are a better fit. In practice you use USE to find the saturated resource behind a service that RED shows is slow.
Ops Singularity's InfraOps and TelemetryOps capabilities capture USE-style resource signals, which Sentinel AI uses to find the saturated resource behind a service incident.
Utilization, Saturation, Errors, the three signals to check for each resource.
Saturation, the queued or waiting work, often rises before utilization maxes out, making it an earlier warning of a bottleneck.
Ops Singularity turns open telemetry into autonomous, governed resolution. See it on your own stack.