From a conversation on Freenode's #LVM

snaphost monitoring - it writes warning to syslog (it is "preview" and must be explicitly activated)" - isn't a warning that must be explicitly activated and then monitored for, and escalated, rather timid for such an important event? Also, does LVM have a way of sending a message to the administrator to warn of the impending overflow?

there must be dmeventd running, then (if configuted in lvm.conf) it monitors all snapshots automatically and prints warning on some % of snapshot fill - to warn before overflill. * you could probably schedule something to use 'lvs'

what if the size of the changes exceeds the size of the snapshot volume? Is there a way to automate the removal of the snapshot once the backup is done?