What it looks like to me is that sickle starts a 20-second timer when it is first run, and whenever it receives a ping it restarts the timer at 10 seconds. If sickle is suspended then after resuming it restarts the timer at 10 seconds. If the timer ever expires then sickle triggers a reboot.
If that is how it works then the patch I posted in
post #13 should be completely safe, as it just increases those timer durations.