Ever tried to track down the time of a reboot for a UNIXoid (Linux, *BSD, …) machine digging through the history-page of the cpu-column of the Xymon webinterface? The Windows-clients for Xymon have a separate column for uptime by default but *nix clients uptime is traditionally reported mixed with the cpu-status since the BigBrother days.

For UNIXoid systems these mix makes searching for boot-events unnecessary complicated as you have to check every yellow status …

… including those caused by high cpu load.1

Since Xymon v4.3.18 it is possible to generate a separate uptime-status column by adding the --uptime-status-option to the invocation of xymond_client in tasks.cfg.

The following snippet from the demo site shows this in full:

    ENVFILE /usr/lib/xymon/server/etc/xymonserver.cfg
    NEEDS xymond
    CMD xymond_channel --channel=client --log=$XYMONSERVERLOGS/clientdata.log xymond_client --uptime-status

man 8 xymond_client (online version) states

Generate a separate "uptime" status column. The uptime is normally just reported in the "cpu" status column, but if you would like a separate status column for the uptime, you can use this option. It is useful if you want to generate an alert in case of a reboot, without having this alert mixed in with the cpu load alerts.

The uptime check in analysis.cfg is unaffected by this change, so configurations like

UP 1h

to go yellow for 1h after bootup or

UP 1h 365d

to go yellow after 1 year of uptime too still work. As expected the cpu-column is not affected by boot-events any more but the new uptime-status changes color.

  1. Of course you could calculate the boot time from the uptime — but that is a bit inconvenient.