[Beds] Access Log Analysis

[David Pashley]

On Jan 30, 2006 at 17:03, Darren Parkinson praised the llamas by saying:
> Hello All,
> 
> Just a quick question regarding access_log analysis.  I've just tried Analog
> and Visitors for analysing the contents of an access_log file and also
> attempted to get Webalizer and Yaala to work too but with no luck.
> 
> Anyway, both that I did get working give slightly different results as to
> the number of visitors to the website.  Does anyone else have any experience
> of using tools such as these and maybe recommend anything else?

I find AWStats(http://www.awstats.org) to work rather well. 
> 
> I wrote some Perl code to take a look through the file too and I worked out
> 116 different IP addresses, but clearly this doesn't necessarily mean that
> this is the number of unique visits, since users may be sharing an IP,
> although I suspect is more accurate in this case given that I'm only looking
> at Sunday data for a non business site... Visitors worked it out as 260
> unique visitors and Analog said about 200.  So, who am I to believe and what
> are your experiences.
> 
I can't remember the exact reason for this, but it's entirely possible
that Analog believes some people are part of a returning session. Maybe
using session urls or something. I know I saw something on the awstats
website about this.

> Finally can anyone explain HTTP 304s?
> 
304 is "not modified". Basically the first time the browser downloads
the page, the server gives it a modified date. The next time the browser
a if-last-modified header. If the page hasn't changed since that date,
the server doesn't bother serving the page and just tells the browser to
use what it had before.


> Thanks.
> Darren.

> _______________________________________________
> Beds mailing list
> Beds at mailman.lug.org.uk
> http://mailman.lug.org.uk/mailman/listinfo/beds

-- 
David Pashley
david at davidpashley.com
Nihil curo de ista tua stulta superstitione.