Howdy,
Can anyone suggest a network monitoring system that knows the difference between a cisco 1701 and a GSR 12810/6500, etc?
What I mean is, many times these days there are several different sub systems you have to monitor inside of a router/switch and not just interface utilization, the "CPU", and the "RAM".
Statistics such as CEF utilization, fabric utilization, PFC/DFC, various line card statistics, etc?
Can anyone recommend anything other than "customize MRTG a lot" that we can use to get a better look into these systems?
thanks,
-Drew
Most of these threads usually result in telling the poster to RTFM with a link to it 
I'm too lazy to link the manual.
c-nsp has extensive archives with lots of questions about various specific SNMP mibs that weren't immediately evident from RTFM.
It all comes down to SNMP to the best of my knowledge.
Drew Weaver wrote:
Ah, I was mainly interested in an Orion like system that actually has all of that kind of worked-in.
Thanks for the heads up.
-Drew
We use Cacti for this purpose, but it still requires creating custom
datasources for the vendor-specific SNMP MIBs.
William
True. While you don't want the MRTG answer, I'd suggest looking at Cacti.
There's a large library of device profiles people have put together so as to
prevent you from having to hunt down MIBs/OIDs for devices. If you have a
database of your devices, it's fairly trivial to import them into Cacti once
you have the device profiles (I use a shell script and curl).
Have you looked at OpenNMS ??
Drew Weaver wrote:
Ah, I was mainly interested in an Orion like system that actually has all of that kind of worked-in.
Yeah I got that. I am not aware of anything that does that. Not to say it doesn't exist, but if it does it's somewhat well hidden.
http://www.frank4dd.com/howto/nagios/cisco-patch-update-monitoring.htm looks interesting and has come up in several searches I've done in the past when needing to monitor cisco kit.
I'm guessing CiscoWorks might have what you are looking for?
I've never been happy with the big commercial NMS products. NAGIOS(with SNMP plugin)+mrtg/cacti+smokeping has served me and many of my colleagues very well.
There is alerting and trending which must be taken into consideration.
Alerting is pretty easy, especially with giving nagios knowledge of hierarchy (if a switch or router stops responding you don't get alerts for all the servers attached/downstream of it). You can easily automate the setup with things like nmap2nagios and other tools.
Trending (which it seems is your primary concern) is harder. Zabbix has some cool SLA reporting and dashboards.
I seem to recall a FLOSS NMS thread a few months ago on here, or maybe it was c-nsp. Dunno.
Are you primarily concerned with monitoring, or with trending/capacity planning?
We use Cacti for this purpose, but it still requires creating custom
datasources for the vendor-specific SNMP MIBs.
+1 for cacti.
I think pretty much everything requires bringing in the mibs and setting up mappings etc.
I've used Nagios/Cacti/Ganglia/MRTG.
If you are interested in an Orion-Like system, but can't foot the bill
for it, maybe look at IpMonitor. Solarwinds acquired IpMonitor a while
back, so their sales reps will try to sell you on Orion.
I've had many years of good luck with it (IpMonitor) and Solarwinds
seems to be handling the software pretty well.
Drew Weaver wrote:
Howdy,
Can anyone suggest a network monitoring system that knows the difference between a cisco 1701 and a GSR 12810/6500, etc?
What I mean is, many times these days there are several different sub systems you have to monitor inside of a router/switch and not just interface utilization, the "CPU", and the "RAM".
Statistics such as CEF utilization, fabric utilization, PFC/DFC, various line card statistics, etc?
Can anyone recommend anything other than "customize MRTG a lot" that we can use to get a better look into these systems?
Netdisco and zabbix both have decent auto-discovery built in.
zabbix will auto build a template for you which you can then deploy to
your devices.