I have a demo version of InterMapper for linux (http://www.dartware.com).
It's probably the best monitoring tool I've ever used, but historically,
it's only been available on Mac, until now.  It's very very fast, and takes
very little resources.  Has auto-discovery, and it just plain works great.
Oh, and it's very cheap compared to most other commercial products.  

I'm currently evaluating some monitoring products which need to keep track
of a few hundred network devices, and several hundred network links
(traffic, interface down, etc..).

Right now, I'm using Nagios (http://www.nagios.org) and Cacti
(http://www.raxnet.net), however, Nagios is a nightmare to administer with
this many devices (edit 3 config files to add or remove a single device),
and Cacti is not really really scalable enough at this point to handle this
many links.  There's a threaded daemon available for Cacti, but it's not
fully mature yet, and the release version of it is very tedious to set up
with this many devices.

Besides device monitoring, I need 2 more things.  I need to be able to keep
historical traffic data for all of our links, and I also need something that
will generate an "executive report" every week which lists the percentage of
network uptime, a listing of outages, and also a list of the links which
exceed an average of 60% utilization any day of the week.

Unfortunately, Intermapper only provides the monitoring portion of this.  It
has some graphing capabilities, but not really a full fledged database of
past data that I can go back and look at like with Cacti or Cricket.  Also,
it doesn't seem to offer any sort of executive reporting option.  I could
write something if I could figure out how to pull info from it, and ideally,
I would like it if I could make intermapper's traffic polling save the data
to an RRD database.  Has anyone tried this?

I'm looking at some other products also, however, they are much more
expensive, and unfortunately, many of them are like 4 different products
hacked together into one web interface, which is fine and dandy, but
typically, at least one of the products hacked in is something that sucks.

Otherwise, does anyone have any recommendations for a decent
monitoring/reporting package?  Free is good, but not a requirement.  Ease of
administration is a must though because there are just too many devices and
links to have to add them using a tedious process.  Also, the monitoring
program must not only check for down devices, but also alert when a device
has a down interface (like a frame link or something).  Both Nagios and
Intermapper do this.

Big Brother is not what I'm looking for.  Nagios is close, except for the
administration part of it.  Cacti comes close for traffic management, and
will probably fit the bill when the new version comes out, but it's not
ready yet.  Using Nagios and Cacti, I could very easily write something in
Perl that would generate a weekly report.  Still kind of a hack, but at
least I would have control over what it does.  But I'm still willing to pay
money if I can find something that does it all well, and is not going to
take up obscene amounts of my time for day to day administration.

Jay

_______________________________________________
TCLUG Mailing List - Minneapolis/St. Paul, Minnesota
http://www.mn-linux.org tclug-list at mn-linux.org
https://mailman.real-time.com/mailman/listinfo/tclug-list