CPU monitoring

Overview

CPU metrics come from the Capture agent running on the monitored host. The agent reports the fields below on every check; Checkmate plots them in the infrastructure monitor's detail view.

Fields reported

• usage_percent — current CPU utilization, 0–100
• free_percent — idle capacity
• physical_core / logical_core — physical and logical core counts
• frequency / current_frequency — base and current clock speed in MHz
• temperature — per-core temperature readings when the host exposes them (Linux sensors, SMART on some platforms)

Temperature is often empty on virtualized hosts and on macOS without extra tooling.

Reading usage

A sustained usage_percent above your configured warning threshold is the signal to scale or investigate. Short spikes to 100% are normal during GC, builds, or batch jobs and don't necessarily mean trouble. Set the warning and critical thresholds in Hardware thresholds.