> Telling them apart from legitimate crashes from remote/noninteractive sessions is a lot more difficult,
Actually, I just realized it's very simple. We can keep crashes from remote/non-graphical sessions and alert about them, but ignore crashes which happened in a previous graphical user session.
> Telling them apart from legitimate crashes from remote/ noninteractive sessions is a lot more difficult,
Actually, I just realized it's very simple. We can keep crashes from remote/ non-graphical sessions and alert about them, but ignore crashes which happened in a previous graphical user session.