Dealing with Packet Loss and problematic nodes

Started by james_ie, May 16, 2017, 11:06:35 AM

Previous topic - Next topic

james_ie

Hello

We have some nodes that for various reasons have packet loss. We end up getting spammed with alarms. Of course these nodes are to be fixed to resolve packet loss.

However in the interim, is there any way to only monitor if these nodes stay down for 10 minutes. Dont really care about any interfaces etc.

So to sum up, monitor certain nodes and only poll them every 10 minutes if they are up or down?

i would really appreciate any assistance on this matter.

thanks very much 

StanHubble

You can change the default polling interval for each of these DCI's or you can set a threshold of "last(10)> {somevalue}" to trigger the event.


james_ie

#2
Hi thanks for the info

I'm talking about not even using the DCI though. i want to use custom defined intervals for "Status Polling" . the default server is 60 seconds, i would like to increase this manually by zone/node to negate the packet loss

Is my only option to create some custom DCI and have it do the status calculation?

Victor Kirhenshtein

Hi,

you can set number of polls required for status change. That way polls will still be every 60 seconds, but status will only change when configured number of consecutive polls fails. You can set that on interface level.

Best regards,
Victor

james_ie

#4
Hi Victor

Thanks so much for the info!
Can it be done on a per node basis instead of interface?

If not is there any other way around this?

thanks again!