Monitoring of Procurves with MRTG

[Neil_77]

Anyone out there using MRTG to monitor their procurve switches (4000M/8000M)?

Ive just set MRTG up for a few of our switches - pretty straightforward so far. And I now have some nice pretty graphs of traffic in and out for each port.

But can anyone tell me how I would go about monitoring some more "advanced" stuff - such as CPU utilisation, discarded packets, errors, that kind of thing.

Thanks,

Neil

[Preston Gallwas]

Hi Neil! To monitor things with MRTG via SNMP, you need to know the OID of what you want to monitor. Currently, I only monitor our cisco layer 3 catalyst 4502 CPU, but heres an example of what I did to MRTG to get it to chart CPU. The OID is a generic CPU mib...im not sure if procurve switches have the same ID for CPU -- look it up in the documentation or use a MIB browser to find it.

Here is a sample of our MRTG config

### Core Catalyst #3
Target[cpu.3]: 1.3.6.1.4.1.9.2.1.58.0&1.3.6.1.4.1.9.2.1.58.0:public@$IPADDRESS
YLegend[cpu.3]: CPU Utilization
ShortLegend[cpu.3]: %
MaxBytes[cpu.3]: 100
Options[cpu.3]: nopercent, gauge, unknaszero, growright
Unscaled[cpu.3]: dwmy
Legend1[cpu.3]: CPU Utilization
Legend2[cpu.3]: .
Legend3[cpu.3]: Max value per interval on graph
Legend4[cpu.3]: .
LegendI[cpu.3]: CPU:
LegendO[cpu.3]: .
Title[cpu.3]: CPU Utilization on $DEVICENAME
PageTop[cpu.3]:

CPU Utilization $DEVICENAME

Colours[cpu.3]: GREEN#00eb0c,BLUE#0000ff,GRAY#AAAAAA,VIOLET#ff00ff
WithPeak[cpu.3]: ymw

[Les Ligetfalvy]

Well.. CPU utilization is pretty easy since it is a range between 0 and 100. What values are you hoping to get for the others?

Totals don't graph well so you would have to look for values per second and such. That is probably best gotten with RMON but I don't have a clue what the 4/8M has in that regard.

PCM+ will give you the RMON stats and does a nice job of consolidating the worst measures onto one screen. It would be nicer if one could recall historical values more that one hour old.