Help me with ping latency trigger

[easy_john]

Hi.

I need to send alert, when all pings to server more then 100ms, last 2 minutes.
But do not need this trigger to fire, when single peak.

I try to create trigger {Template_Host:icmppingsec.avg(120)}>100, but it fires with some like this in history:

2008.Jan.29 15:44:08 0.7300
2008.Jan.29 15:43:37 509.0000
2008.Jan.29 15:43:02 0.5200
2008.Jan.29 15:42:31 1.0200

How create trigger, than fire only with long time latency peaks.

[easy_john]

nobody knows?

[cstackpole]

OK I am not 100% on this, but I am more then willing to take a stab at it...


{Template_Host:icmppingsec.avg(120)}>100&{Template _Host:icmppingsec.delta(120)}<100

That should filter out that one really high value...

Hope that helps (or at leasts sparks interest from others to join in and post)
Have Fun!

[easy_john]

OK I am not 100% on this, but I am more then willing to take a stab at it...
{Template_Host:icmppingsec.avg(120)}>100&{Template _Host:icmppingsec.delta(120)}<100
That should filter out that one really high value...

at this time i check with "{Tpl_Host:icmppingsec.min(60)}>100" not best solution, but trigger work, when link has overload or bad connection...

[tighep]

Hi.

I need to send alert, when all pings to server more then 100ms, last 2 minutes.
But do not need this trigger to fire, when single peak.

I try to create trigger {Template_Host:icmppingsec.avg(120)}>100, but it fires with some like this in history:

2008.Jan.29 15:44:08 0.7300
2008.Jan.29 15:43:37 509.0000
2008.Jan.29 15:43:02 0.5200
2008.Jan.29 15:42:31 1.0200

How create trigger, than fire only with long time latency peaks.

My understanding is that you're taking an average of the ping times for the last 120 seconds. In the example you gave the average is over 100. You would either need to increase the frequency of the check to something other than 30 seconds, or raise the average that you're looking for. Unfortunately I can only help with why you saw the event, and not how to do what you wanted.

[easy_john]

My understanding is that you're taking an average of the ping times for the last 120 seconds. In the example you gave the average is over 100. You would either need to increase the frequency of the check to something other than 30 seconds, or raise the average that you're looking for. Unfortunately I can only help with why you saw the event, and not how to do what you wanted.

no, average not allow, it generate event for single peak, and I do not want to receive message on every single peak, only while channel for some long time lose speed!