[Bayonne-devel] Monitoring Bayonne traffic with MRTG...

[David Sugar]

It seems me very natural to use MRTG to track call traffic trends and 
history from a Bayonne server.  I am curious if anyone else has done 
this already.  In fact, one can scan the testing branch usage file or 
create a graph for individual trunk groups from the stats file with 
relativily little work.  This works great if mrtg is ran from the same 
server Bayonne is on since one can write a simple perl or shell script 
that is invoked from the mrtg cfg file and have it return the peek 
inbound/outbound traffic port usage from the Bayonne calls and usage 
files, and these stats are rolled over on a fixed interval by Bayonne 
already.

The one problem comes in when trying to monitor traffic from remote 
Bayonne servers.  We don't currently publish them network visible as 
part of the bayonne shared net traffic.  Perhaps we should.  Another 
solution might be to adapt a shell script to attach to a tcp port using 
Mark's excellant tcp console monitor.  A third possibility I had 
considered was to write an inetd deamon that could be queried remotely.  
I am curious how people may feel about either of these methods, 
especially those who might already be using mrtg with Bayonne.