On Tue, 28 Mar 2000, Brad Groshok wrote:
> Looking at MRTG stats for traffic to/from our squid box...
> The traffic coming in to the squid box is almost the same as the
> traffic going out of the box (to our clients)
"Why is my cache's inbound traffic equal to the outbound traffic?"
http://www.squid-cache.org/Doc/FAQ/FAQ-12.html#ss12.22
> Not sure if I'm expecting too much from squid, but I would have
> expected there to be a greater diff in incoming vs outgoing
> traffic.
MRTG monitoring of a physical interface (or even a virtual subinterface) is
*useless* for determining the effectiveness of Squid (or any cache for that
matter); you will need to measure either using a more intelligent network
layer tool (such as Cisco NetFlow, though you will have to perform a lot of
postprocessing on the data to extract useful information), or move your
measurement up to the application itself - either use MRTG to monitor Squid
directly through Squid's SNMP support, or postprocess Squid's access logs in
order to discover the effectiveness of the cache.
cachemgr.cgi can give you an ongoing snapshot of the cache's usefulness, but
the numbers can vary quite dramatically from snapshot to snapshot so choosing
a fixed time period (eg., 24 hours - 1 day) and postprocessing the access logs
is going to be the best way to gauge cache effectiveness from day to day.
Cheers..
dave
Received on Wed Mar 29 2000 - 09:09:42 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:52:28 MST