[Discuss-gnuradio] More CloudFlare statistics

[Johnathan Corgan]

So, after a week using CloudFlare to cache gnuradio.org website
content, we have some interesting statistics.

In the last seven days, there have been over a million hits on
gnuradio.org, almost exactly half of which were users and half were
web crawling robots.  The non-crawler traffic came from 11,426
distinct IP addresses, using a total of 54 GB of bandwidth.

If this seems high, it is.  The large majority of of this traffic is
git traffic over http, which can generate hundreds of hits for a full
git clone operation.  Also, we don't attempt to exclude any web
crawlers from scraping the site, so pretty much every GNU Radio source
code file from every revision in the repository gets indexed.  Several
times, in fact--through the dynamically generated pages in Redmine,
through the git HTTP interface, and through the cgit repository
viewer.

One of the capabilties of CloudFlare is to insert Google Analytics
javascript in each page that looks like it is being loaded by a
browser (and thus not git or webcrawlers).  So for users that have
Javascript enabled, we've been seeing much more typical numbers
consistent with a project our size.  For the last week, then, there
have been:

- 38,514 page hits, from
- 10,494 visits (thus 3.85 pages per visit), from
- 5,924 unique visitors, from
- 113 different countries

I'm attempting to get accurate population counts for each country so
we can see which have the highest per-capita interest in GNU Radio,
but that could take a while.  Any one want to guess which ones that
will be?

Anyway, fun stuff!

-- 
Johnathan Corgan
Corgan Labs - SDR Training and Development Services
http://corganlabs.com