Wish: Multiple instances of alerts on the same service/host
Sam Stickland
sam_mailinglists at spacething.org
Wed Mar 21 12:53:23 CET 2007
Ståle Askerød Johansen wrote:
> (This may appear twice. I fumbled with my subscription confirmation)
>
>
> Here at the University of Oslo we are currently running Nagios
> alongside our current monitoring system in order to check if
> Nagios suits our needs.
>
> So far, we are very happy with most of what we see. However, we
> also consider using Nagios (with some suitable www-interface) as
> our primary alarm console. This means that we will want to feed lots
> of passive checks into Nagios from several other systems.
>
> Let me give you an example:
>
> - we want to forward SNMP-traps to Nagios from the management cards of
> our Dell and HP servers.
> - we setup our trap-receivers to submit this through NSCA.
> - on the nagios server, we define the service "snmp trap" on all the
> relevant hosts. the service is volatile and not active.
> - we test.
> - the hardware sends for instance "Fan 2 not OK". Nagios receives this
> as a critical event. let's pretend the operator uses some time to fix this.
> - in the mean time, the hardware on the same host sends for instance
> "battery needs replacement". Nagios receives this as a critical event,
> but the previous event if NO LONGER visible in the interface.
>
> Some may argue that we need to make separate services for each type of
> trap we want to receive, but sheer numbers make this not very elegant.
>
Just for reference, this is how HP Openview NNM handles this. It scans
the MIB files and generates an NNM event for every type of trap that
can be received. Perhaps there could be some milage with this approach?
Of course, this doesn't solve the multiple traps of the same type
problem, where just the parameters differ (e.g. BGP transition changes
for different peers).
S
> We need a way to tell Nagios that "this service is of a special kind
> whose events should not replace each other as they are received". This
> will make it easier to use Nagios and a suitable web-gui as a central
> alarm receiver without adding thousands of new services.
>
> The same problem also makes it difficult to make, for instance, a plugin
> that monitors all userdisks on a host and reports to a service
> "userdisks", since the events will overwrite each other.
>
> Has anyone else thought of this? Is it difficult to implement? Are we
> wrong in assuming that this is impossible with the present Nagios? Have
> we misunderstood completely? Is it a stupid and childish idea? :-)
>
>
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
More information about the Developers
mailing list