Smart Alert?

[Noman0118]

Hello,

I want to know how to do the below following, i have been searching ways to do this but have yet to find anything solid. Please, help.

A alert can be created for a ping check to google and if that check fails, run a control action to disable the alerting for the monitoring system itself.

[admin]

What do you mean by that? Do you mean if Google is down, you want to
disable alerting for HQ? There is no such action available. You
could potentially write an HQU that allows you to do that, and then
create a control action for the alert (HQ Enterprise feature) to
invoke a script that calls back into HQ through HQU to disable alerting.

Charles

[Noman0118]

the end goal is to implement notification suppression mechanism. I want to check internet connectivity by pinging google.com or some major site, if i cant get to it, it means the internet connection link (most likely) is down. Then i want to disable alerting on HQ entirely, the only alert should be about the internet link being down. We have the free free in production as of now and I am literary the one-man team to do this. I am lost and we have very limited budget and resources. Any help on how to EXACTLY implement this would be great. I cant imagine this feature missing in Hyperic.


Thanks in advance!

[admin]

There is a plugin on hquplugins.org which shows you how to globally
disable alerts from a shell-script (nolert). You can use this in
combination with a Script service to disable alerts when your
threshold is reached.

-- Jon


On Jun 17, 2008, at 8:43 AM, noman wrote:

> the end goal is to implement notification suppression mechanism. I
> want to check internet connectivity by pinging google.com or some
> major site, if i cant get to it, it means the internet connection
> link (most likely) is down. Then i want to disable alerting on HQ
> entirely, the only alert should be about the internet link being
> down. We have the free free in production as of now and I am
> literary the one-man team to do this. I am lost and we have very
> limited budget and resources. Any help on how to EXACTLY implement
> this would be great. I cant imagine this feature missing in Hyperic.
>
>
> Thanks in advance!

[Noman0118]

Jon,

This looks very promising, i will try this out and get back to you soon. Thanks again for all your help.

Best Regards,

[Noman0118]

I downloaded the the plugin you mentioned and manually tested it. I am confused how to implement it. I am new to this, can you please explain in details.

Thanks for your help!

[deeboh]

Hey Echo5. When creating alerts, in the condition set area there is a check box at the bottom of the definitions page which disable alerts once they are fired, until they are fixed.

here is what the text says...

Generate one alert and then disable alert definition until fixed


Is this what you're looking for?

Best,

Curtis

[Noman0118]

Hey Curtis,

Sorry i am not looking for that. I have found and installed the HQplugin script that when run can disable or enable alerts on the global level on the HQ server (refer to earlier posts). I basically want to ping a major site (google.com or cnn.com, etc) if i cant reach to it MOST likely the internet link is down (that has been the case for us several times, courtesy of comcast :(.. ) and i want to run the scrip to disable ALL alerts on the HQ server except the Internet is down one. Basically, implement a poor man's alert suppression mechanism. Any thoughts about this?

~noman

[deeboh]

Ah haaaa. Now I see your dilemma. the HQ Plugin stuff appears to be your best bet. I'm a much better shell scripter so my approach might not be to your tastes. However, i've had some success with traditional control plugins (Pre HQ 3.1) that programatically shutdown server processes based on a return value being interpreted by an xml-plugin.

pseudo script:
poll comcast (using monitor interval)
if up
continue to poll
don't change monitoring return value
{comacast in sequence responses within acceptable time
return = 0(good)
}
if down
continue to poll
change monitoring return value
{comacast returned nothing or out of sequence response
return = 1(bad)
}

this way you can create your alerts based on the return value returned from your script. Also with the return values and case statements you can return values that show shades of gray rather than a hard up or down. So you can get a fair warning if the link(s) are slow or going down or in trouble.

with a properly constructed xml-plugin you can even use standard hyperic control values for http monitored entities, processes execute another script whatever.

Then copy the script,xml-plugin pair to each of your servers and away you go.

However I believe the HQ Plugins sound like a more sustainable direction.

I hope this helps,

Best...