Thanks a lot! Options: --debug --cfid was very useful. I had found at least one place, which most probably caused the problem.
I would like to explain a bit:
In alerts config. I had used some variables like:
$HOSTS_PROD=host1,host2,host3
$HOSTS_TEST=thost1,thost2
$HOSTS_DEV=dhost1
And some rules, including for $HOSTS_DEV (some separate as well as some together with $HOSTS_TEST)
and merged variables too:
$HOSTS_ALL=$HOSTS_PROD,$HOSTS_TEST,$HOSTS_DEV
This worked fine until we switched of dhost1 host (reason does not matter) and I had commended variabled definition:
# $HOSTS_DEV=dhost1
But rules remained. As result there was rules for empty element
# development
HOST=$HOSTS_DEV SERVICES=$SVC_DB_BUSN COLOR=red,yellow,purple
MAIL=$CM_SUPPORT_DBA DURATION>15m REPEAT=60m RECOVERED DURATION<180m FORMAT=PLAIN
this rule really resulted to:
HOST= SERVICES=dbinvobj COLOR=red,yellow,purple
MAIL=user-67221db774f9@xymon.invalid DURATION>15m REPEAT=60m RECOVERED DURATION<180m FORMAT=PLAIN
This rule most probably worked out in some cases together with correct one.
Best regards,
Andrey Chervonets
SIA CoMinder
http://www.cominder.eu/