item1
key_ :proc.num[ntpd]
interval:30

item2:
key_ :vfs.file.exist[/tmp/maintenance]
interval:30

trigger
Expression:{TEST_SERVER:proc.num[ntpd].count(#2,1,lt)}=2 & {TEST_SERVER:vfs.file.exist[/tmp/maintenance].last(0)}=0
Event generation:Normal

When doing setting like the above, an event is notified twice, and there is a possibility that wrong evaluation is done.
When the same trigger is estimated by HistorySyncer or Timer twice at the same time, this problem should happen.

The export file made for a test is attached.
If the problem is not reproduced, please copy the host. It should be easier to reproduce.