Reccomendations on enterprise monitoring tools
Spin
I'm looking for some recommendations on enterprise monitoring tools. What's
your favorite enterprise monitoring tool or utility (such as NetIQ, HP
OpenView, Microsoft Operations Manager, What's Up Gold or Open source such
as Naganos) and why (ease of use, ease of installation, low cost)?
--
Spin
Damon Getsman
or WAN.
* Ntop- for extremely detailed network usage and monitoring of
precisely what kind of traffic and where the traffic is going.
* MRTG- for bandwidth graphing on multiple interfaces so that you can
see if something bogus is going on.
At least that's what works pretty well for our WAN. :) I'm open to
other recommendations, too, however, if anyone has other favorites.
Ntop was the only one of those that was really easy to set up; the
other ones monitor all kinds of different stats over SNMP, and nagios
also utilizies ping, ssh, and a few other tools to keep watch on
various aspects of the hosts and multiple devices that it's hooked up
to.
http://www.lookupanyone.com/namelistings/damon-getsman-andrew-getson.html
Damon Getsman
ITrx
Martijn Lievaart
I have experience with Nagios and Cacti and some with Zabbix.
Nagios is very good. I personally find it extremely easy to set up, but I
heard others bitching about it, YMMV. A particular strong point is that
it is relatively easy to create custom plugins. For big installations
tuning takes some black magic, but its not that hard. Basically, when it
works, it just works. And because there is no database, it is easy to
interface, just some scripting required.
Cacti is nice. But I'll avoid it in the future because of a couple of
annoying things in actual deployment. It can easily overload your
machine. If it crashes or hangs, which happens, deep magic is required to
repair it. Its hard to interface to other stuff. But most importantly,
there are bugs that make the database accumulate cruft over time that
cannot be removed, at least not that I could find without digging very
deep. And that cruft does affect performance as you will be monitoring
data sources that don exist anymore, which really slows it down.
Zabbix is the up and coming kid. I hear very nice things about it and we
actually run it on a couple of machines. Other people set it up, so I
cannot really comment on it, but be sure to give it a look.
I was extremely disappointed with what's up gold. Spend your money
elsewhere, the open source tools are way better.
Anyway, if you plan on a big installation, get some big iron. Any recent
server will do, but don't install on some old machine left over. If you
go really big (say over 500 nodes to check) go for distributed monitoring.
M4