Observable Metrics

Metrics describe the measurable activity of processes and user sessions. Each metric supplies a Boolean value to a watch expression. Thresholds determine when measured activity counts as usage; the expression determines which combinations keep the resource active.

• Metrics
• Watch expressions

Thresholds

If not specified otherwise, the thresholds apply to the combined counters of all matching processes. On a <Process> element, watchChildren="true" includes child processes. Using it on a session selector it applies to all of the session’s processes.

Inspection interval

Thresholds are evaluated over the inspection interval set by SystemMonitor timeout. A rate such as 1MB/s is an average over that interval. An absolute value such as 10ms or 100kb applies to the amount observed during the interval.

Calibration

The measured metric values are usually equivalent to those, that your platform’s Task-Manager or Activity Monitor displays.

Setting any threshold (like minCPU, minGPU, minIO, or minTraffic) to 0 does not constrain the detected usage of a resource, but enables the measurement of the respective metric, which value you can then read via the log.

All configured threasholds are measured, including those not referenced by any expression. Using an unsupported or unknown metric is not allowed. If no monitored process can supply a watched metric during inspection, an error is reported instead of assuming activity or inactivity. See Troubleshooting.