[mrtg] Re: Combining Graphs

[Mike Starkweather]

Well, I guess you could write a script to do the polling and then format the
individual files for MRTG the way you want them.  Read up on external files
in the Target statement.  Poll the box once, put out individual files per
device and add them to get the overall figure.  I have been doing this for
more than a year on a pool of 22 Shiva devices.

I'm not sure how you would handle the case of the missing devices (down for
maintenance).  Maybe some smart polling routine that, once it senses a dead
device only tries it once an hour or so.  You want to avoid manual
configuration changes if at all possible.

Mike Starkweather
-----Original Message-----
From: Denis Saveliev <Denis.Saveliev at theplanet.net>
To: 'mrtg at list.ee.ethz.ch' <mrtg at list.ee.ethz.ch>
Date: Friday, January 14, 2000 5:54 AM
Subject: [mrtg] Re: Combining Graphs


>
>> Don't know if this is what your looking for or not, but we have (3)
>asymmetric
>> serial lines on a Cisco that are load shared.  We combine the ifOutOctets
>and the ifInOctets
>> values for the three into one graph like so (for the input side)
>
>> Target[name]: ifInOctets.8&ifInOctets.8:community at router +
>>          ifInOctets.11&ifInOctets.11:community at router +
>>          ifInOctets.12&ifInOctets.12:community at router +
>> MaxBytes[name]:  <size of three lines combined>
>
>This works ok if you have a few routers.... But in our network we have 450+
>Cisco as5300's. Well, I guess we probably can put them all into one
Target[]
>string and it would probably work (don't know if there's a string length
>limit in mrtg) but:
>
>1. Since we need to monitor the number of users online on every 5300 as
well
>total number of users mrtg have to query every router twice - once to get
>number of users for the stats for the router and once to get a summary
>accross the dial pool leading to double load on the machine running mrtg
and
>double traffic across the network.
>
>2. Mentioned before is inconvinient but we could live with it. What more
>important is that if any of modem racks is down for maintenance it affects
>the whole stats since mrtg cannot get the responce for the snmp query and
>therefore failes to calculate ... + ... + ... + ... equasion which results
>the flat summary graph!!! As modems can be down for days this is totally
>unacceptable (well, I guess we can keep editing mtrg configs but if modem
>goes down in the midnight we still got flat graph till next morning).
>
>I believe it would be really useful if there was a way in mrtg to get the
>summary stats across the number of different routers. Something like:
>
>Target[name1]: string:community at router1
>Group[name1]: DialPool1
>
>Target[name2]: string:community at router2
>Group[name2]: DialPool1
>
>Target[groupname]: Group:DialPool1
>
>Regards,
>Denis
>--
>Planet Online Network Engineering
>
>--
>Unsubscribe mailto:mrtg-request at list.ee.ethz.ch?subject=unsubscribe
>Help        mailto:mrtg-request at list.ee.ethz.ch?subject=help
>Archive     http://www.ee.ethz.ch/~slist/mrtg
>

--
Unsubscribe mailto:mrtg-request at list.ee.ethz.ch?subject=unsubscribe
Help        mailto:mrtg-request at list.ee.ethz.ch?subject=help
Archive     http://www.ee.ethz.ch/~slist/mrtg