After a system has been rebooted there may be many clients that are waiting for service. If they all connect at once then they can cause a temporary load spike that can trigger the load average checks - which then means another reboot... I think it would be good to have an option to allow a delay after watchdog startup before the load average is checked.