Unexpected downtime sucks, usually due to some stupid assumption you made and lack of testing. You just have to hope that the other nodes dont close channels out right away.