It really depends on what you mean by "monitoring" and what you want to monitor.
MRTG was originally meant for recording and graphing the traffic levels on router interfaces via snmp, so it comes with tools to make that happen pretty easily, but in order to graph any other data, you have to put in more effort to obtain the data (figure out the SNMP OID or write a script to get it some other way).
MRTG wasn't particularly efficient when it came to data collection, storage, or graphing, and the person who wrote it supplemented it with RRDTool, which is not only more efficient but also a great deal more flexible (and more complicated).
If what you really meant by monitoring wasn't "collecting performance data", then maybe you meant "sending alerts when something bad happens".
If that's what you meant, then Big Brother, Big Sister, Spong, and (my favorite) Nagios (formerly called NetSaint) are what you want.
There are lots of others - these are the ones I have experience with. Nagios is definitely my recommendation out of that group - it's got great documentation, a very active community behind it, and a very nice interface, with good support for intelligently managing alerts, downtime, outages, etc. It's also far more capable than the others (scalable, distributed, fail-over ready).
http://www.nagios.orgThere's also RedHat's newly updated Redhat Network, which now has an integrated monitoring tool if you choose to purchase it. All I've seen are the screenshots.
http://www.redhat.com/software/rhen/system_mgmt/Those are the inexpensive ones I know about. If you are willing to spend lots more money for slightly less admin overhead, then Computer Associates Unicenter and IBM's Tivoli are both available on and for Redhat too. Openview also has some Linux monitoring capabilities. Unicenter doesn't have an agent (so far as I can find) for Linux, even though their management platform runs there.
Hockey PUX?
