[c-nsp] service monitoring on a small scale?

[Roland Dobbins]

On Sep 27, 2007, at 12:58 AM, neal rauhauser wrote:

>   Looking forward to much wise advise on this point ...

Changes in bps and pps over links which exceed certain thresholds can  
be set up in most SNMP management systems, I believe; this can be a  
good indicator that something's going on which ought to be  
investigated.  Using NetFlow to watch for these same types of changes  
can also be helpful.  A bit of scripting with Nagios, nfdump/nfsen,  
et. al would probably be an inexpensive way to make use of these  
telemetry sources in such a manner.

There's a feature built into IOS called IP SLA which allows routers  
to be configured as probes which can check application availability,  
end-to-end latency and jitter, etc.  The statistics generated by IP  
SLA are accessible via SNMP:

<http://www.cisco.com/en/US/products/ps6602/ 
products_ios_protocol_group_home.html>

There's another IOS feature called EEM which basically allows tcl  
scripting on the router itself, which has obvious applications:

<http://www.cisco.com/en/US/products/ps6815/ 
products_ios_protocol_group_home.html>

Here are some example EEM scripts:

<http://forums.cisco.com/eforum/servlet/EEM?page=main>

-----------------------------------------------------------------------
Roland Dobbins <rdobbins at cisco.com> // 408.527.6376 voice

	   I don't sound like nobody.

                -- Elvis Presley