<div>
    Shows the CPU usage of each node's machine, for example <code>37%</code>. The usage is measured
    on the node itself each time the monitor refreshes, as the mean of three one-second samples, so
    a single short spike does not decide the value. It is still a snapshot and not an average over
    the whole interval. Usage at or above a configurable percentage (default 99%) is
    shown in Jenkins' warning style.
    <p>
        Hover over the value to see more detail, for example
        <code>System: 37% (peak 91%) • JVM: 2%</code>:
    </p>
    <ul>
        <li><em>System</em>: the mean CPU usage of the whole machine, all processes included. The
            highest of the three samples is shown as the peak when it differs from the mean.</li>
        <li><em>JVM</em>: the share used by the Jenkins agent process itself. Build steps run as
            separate processes and are not included, so a high system value with a low JVM value
            means builds or other programs are using the CPU.</li>
    </ul>
    <p>
        Inside a container, the usage is relative to the container's CPU limit. If the Java
        virtual machine cannot measure the usage,
        <code>N/A</code> is shown.
    </p>
</div>