[mrtg] General monitoring question
s.shipway at auckland.ac.nz
Thu May 17 01:01:01 CEST 2007
MRTG is really a monitoring-logging program. In other words, it will
graph a metric over time. Although you can put thresholds in there and
trigger execution, this is a bolt-on and not ideal. Some frontends (eg
routers2) will display a grey bar on the graph when an interface or
device is down (data unavailable rather than zero) but this does not
trigger a proactive alert.
What it sounds like you need is a monitoring-alerting program. IE,
check if something is OK/Warning/Critical and send out some sort of
alert if it is not OK. For this, we use Nagios (free!) but other things
like Tivoli exist to do the same for lots of money.
If you use Nagios, it is nice because both Nagios and MRTG can query
hosts via the same agents, and you can register a URL against a Nagios
host to provide you embedded links to the appropriate graphs. Also, if
you use the routers2 frontend, this has a Nagios plugin to embed Nagios
status pages into the MRTG frames.
Second I was testing one of my sites that is connected, but no one is on
it yet, and I send a large number of pings. I then disconnected the T-1
and the only thing that I noticed monitoring the core router was the
traffic dropped. I could not tell the interface was down. Is there a
way to tell if the interface is up or down with this program?
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the mrtg