A system resets repeatedly because its watchdog is serviced by a periodic interrupt even when the main control task has stopped progressing. Which change would make the watchdog more useful for detecting this failure?