Re: [hobbit] Feature Request

["Craig Cook" <craig (at) cookitservices.com>]

It has been a while since I played with Network Node Manager but it had a 
concept of a "reset level".

eg.
Low point: 10    (yellow)
High point: 20       (red)
Reset yellow level: 5
reset red level:    15

ie. temp gets to 11, we turn to yellow.
temp gets to 20, we turn red
temp drops to 17, we are still red.
temp drops to 15, we turn yellow
back to 17, still yellow
etc.

This helps with the flapping issues...

Nagios has "flapping detection" built into it.  I think it checks for a flapping alarm and disables it for a while (it marks the display that the alarm is flapping).  Not sure the exact logic but I could look and try and work it out.

Craig

----- Original Message -----
From: "Henrik Stoerner" <henrik (at) hswn.dk>
To: hobbit (at) hswn.dk
Subject: Re: [hobbit] Feature Request
Date: Wed, 12 Jan 2005 00:20:07 +0100

> 
> On Tue, Jan 11, 2005 at 03:11:49PM -0600, Daniel J McDonald wrote:
> > My biggest gripe with BB - pagelevelsrecovered...
> >
> > I'd like to be able to page on red and send a recovered notice on green.
> >
> > Yellow is no-man's land - it neither causes a page, nor is it sufficient
> > to send a "recovered" page.
> >
> > That's really important when I'm monitoring temperature.  I use
> > bbmrtg.pl to grab the stats.  In one comm room, for example, I have
> > yellow as 65:75, and red as 60:80
> > With bb, if the temperature starts wavering near 79-80, then I get paged
> > repeatedly about it being too hot, no it's OK, no wait it's too
> > hot...
> 
> I think you've mentioned this before, and now I took a closer look.
> It is certainly possible, but I want to hear what you think of this.
> 
> Alert- and recovery-notices initially get triggered by a change of
> color. So I group the possible colors into three groups:
> 
> 1) Those that are critical (red, purple)
> 2) Those that are OK (green, blue)
> 3) The others (yellow, clear)
> 
> If the color changes from OK -> critical, an alert goes out.
> If it changes from critical -> OK, a recovery message is sent.
> 
> If it changes from a color in group 3) to a critical color, an alert
> is generated, but ONLY if there is no alert active already.  So you'll
> get an alert when it goes green->yellow->red, but if it flip-flops
> between yellow and red you will NOT get any additional alerts.
> 
> Likewise if the change is from a group 3) color to OK. A recovery
> message is generated, but ONLY if there was an active alert. So
> the recovery message is triggered if it goes red->yellow->green, but
> if it just flip-flops between yellow and green, a recovery message
> will NOT trigger.
> 
> I think that is the optimal solution ?
> 
> 
> Henrik
> 
> To unsubscribe from the hobbit list, send an e-mail to
> hobbit-unsubscribe (at) hswn.dk