Alerts & Notifications when Bandwidth flatlines

[mberkley]

So over the past week we have had a issue with one our ISP dropping our data connection into our data center.
Naturally a situation like this causes nothing short of havoc within our environment and long bridge calls. Unfortunately we don't notice it immediately, until applications timeout limit expires and triggers alerts.

Since we are monitoring these pipes in Nagios, can we create a INVERSED alert, where if the bandwidth reaches zero (which it should never) it would trigger an alert? I need to know immediately if these critical pipes go flat.

I simple can't see a method where I can make this a possibility.

[sreinhardt]

You absolutely can do this! You should be able to handle this with check_mrtgtraff as the default bandwidth checks do when looking at mrtg's rrds already. If the thresholds for check_mrtgtraff wont allow inverse warning\critical values, you can setup a standard low threshold check_mrtgtraff check, and use the negate plugin to make it so when check_mrtgtraff reports critical for being OVER the threshold of bytes per second, instead you get an OK stating that it correctly should be over this threshold. This may seem a little confusing, so feel free to ask questions, otherwise here's a doc on negate that might help!
http://assets.nagios.com/downloads/nagi ... Plugin.pdf