Did you see the "Require any condition" / "Require all conditions" dropdown on top? ;-)

Tom

On 09/11/2017 14:48, Laszlo Nemeth wrote:

I used this until now:

ifInOctets_perc ge 20
ifOutOctets_perc ge 20
ifInErrors_rate gt 1
ifOutErrors_rate gt 1
ifOperStatus notequals up

looks like the relation is OR between the lines . I added the "ifInOctets_perc lt 100" and "ifInOctets_perc lt 100" lines in this alert and i got continuously alerts because the percentage are always under 100% :)

I made a mistake or the realtion is OR instead of AND ?

Laszlo

On 2017-11-08 14:17, Adam Armstrong wrote:

You just put two lines in the alert definition.

blah gt 80

blah lt 100

Adam Armstrong

CEO & Lead Architect

Observium Limited

On 2017-11-08 11:00:35, Laszlo Nemeth <lnemeth@emarsys.com> wrote:

Hi

We have the same problem with Nexus 5K and 7.1(4)N1(1).

Is there any option to set up an alerts with a from.. and to.. values? Like if the value greather than 80% but lower than 100% send the alarm? In this case this percentage problem doesn't appear. Normally the alert just run if the percentage is over 80. Can we use an "AND" tag somehow?

Thanks

Laszlo

On 2017-11-08 08:32, Markus Klock wrote:

My Guess is that you are polling a Cisco Nexus which have a broken SNMP-stack, it sometimes reports crazy values for a couple of seconds.

If you are not already running it, try uppgrade to nx-os 7.x

Otherwise you can work around the problem by reconfigure you alert check with "delay 2" which will cause it to not send an alert unless the value is above the threshold for 2 polls in a row.

/Markus

Den 8 nov. 2017 08:15 skrev "Edvinas K" <edvinas.email@gmail.com>:

Hello,

What could be the reason of the generation of false alarm:

	ALERT	Modify
Alert		MoreThan80PortUtilization
Entity		Ethernet1/51
Descr		NXOS1-ITXx2
Conditions		ifOutOctets_perc ge 80 (395649)
Metrics		ifInOctets_perc = 11 ifOutOctets_perc = 395649
Duration		10m 8s (2017-11-08 02:18:10)

As you coud notice the ifOutOctets_perc = 395649  shows some irregular value, which's not on the graphs itself, but the alarm is generated.

During migration the migration period from old observium server to new one, at same time we have two observiums running, so older one have the same alerts configured - but that (older one) haven't generated any false alarm.

So - waht could be the reason ?

Thanks!

_______________________________________________
observium mailing list
observium@observium.org
http://postman.memetic.org/cgi-bin/mailman/listinfo/observium

_______________________________________________
observium mailing list
observium@observium.org
http://postman.memetic.org/cgi-bin/mailman/listinfo/observium