triggers of operation status come too late

[daniel_psx]

hi !!


I have a routers monitoring with zabbix, and included a item
named IfOperstatus1 (1 by port of the router). Well
mi trigger is

{R-CARBONATO:ifOperStatus2.last(0)}#1

1 = operative

but i lost conection las night but zabbix send me 10 hours late a mail
with the problem

i`changed for this now

{R-CARBONATO:ifOperStatus2.avg(60)}#1

it`s correct this parameter ?

Thank!!!


Daniel_PSX

[Nate Bell]

Just trying to sort some things out...

How often is that item checked? What did the data for the item look like when it changed? What did Zabbix's event/actions pages show? (ie. when did Zabbix's trigger go off, and when was an email sent by Zabbix)

Nate

[daniel_psx]

well ... i used WhatsUp Gold software for view the principal interface (ip is necessary), and check the MIB for this interface an copy the octect mib
IfOperStaus.
In zabbix created iten, trigger and action mail to network adminstrators

Severity : Warning

Trigger--- Network administrators ---ON---30 ---ifOperStatus2 se ha perdido conexion con {HOSTNAME}

other tip... the status when the trigger is on is in this case UNKNOWN
last value when disconnected is 3

[lreadshaw]

Just for clarification, did you lose connection to the router completely or did a monitored interface on the router go down?

If you lost connection completely, Zabbix would be unable to talk to the router using snmp and wouldn't get any value returned. Checking the ifOperStatus for a value that equals inoperative won't work if you lose connection to the router completely. In that case, I would use a standard ping check. Or you could use nodata() in your trigger so the trigger is ON when Zabbix gets no value back from the SNMP query. Do a forum search for nodata() if you need more info.

Did that help?