VM unexpected status change notifications

There is some slight disadvantage in having the notification system but not having the VM status notifications there.

For example:

1.We have a running VM with the last transaction: start VM
2.VM is being killed at some point in time by oom killer or some other factors
3.We receive status change for this VM from SNMP, VM changes the status to off, but this event remains silent until cloud's customers see that something is wrong with VM

Instead, we need to have this flow to be something like this:

1.We have Running VM with the last transaction: start VM
2.VM is being killed at some point in time by oom killer or some other factors
3.We receive status change for this VM from SNMP, change the status to off, check that last transaction wasn't stop VM or there was no recent chain of transactions with stop VM in it.
4.We send a notification to the system administrator using a specific configurable notification channel in our /notifications