How IISview counts hits
- Hit
- A record in the web log (successful or not)
- File
- A record in the web log indicating a successful fetch (status
code 200). (So files downloaded, not just checked against a
cache).
- Page
- Page is container document (e.g. HTML, PHP, ASP). A hit
includes
other content such as images, flash and javascript files. These
included files are not counted as pages, only the container
documents.
- Unique pages
- Unique pages are the "File" with any query stripped off. The
number of distinct names (as distinct from number of hits). This
is a
clue as to how much of you site is being reached.
- File type
- The filename extension from each hit's filename after any
query
string or anchor tag have been stripped.
- KBytes
- KBytes of traffic sent. IIS records the number of bytes sent
back
to the browser for each hit. (sc-bytes).
- Unique/Distinct Vistors
- A simple visitor is just a logged IP address. So users
accessing
via NAT will only get counted once. (e.g. 2 different PCs on a
home
network with one IP address on the Internet will be counted as
one. If
the home router reconnects with different IP addresses, these
will
count as different visitors.)
- Session
- A session is a bit more complicated. It is is a unique (IP
address +
User agent string) combination, having a hit at least once every
20
minutes. IISView records the time and file when a combination is
first
seen. The time and file of this combination is tracked until
there has
been no requests for 20 minutes. The time between the first and
last
contact is the session time. Mean and standard deviation
statistics are
collected for the number of fetches and length of stay for each
classification of stay.
- Concurrent Visitors
- Maximum concurrent visitor sessions.
- Top Reader
- Top readers are based on the top visitor IP addresses, and
then
the corresponding reverse lookup giving the hostname.
- Response time (mS)
- IIS reports the time taken for each hit. This response time is
averaged (and peak measured) by day and by
timeslot. So daily response time is the sum of all the response
times
for the day divided by the numner of hits on that day. Timeslot
is the
sum over the month of all the response times in that hour (e.g.
13:00->14:00) divided by the number of hits in that timeslot.
- Country/Domain
- Last field of the hostname
- Company/ISP
- Generally last 2 fields of fully qualified hostname indicate
the
company or ISP for US based domains under .com and .net.
Correction is
made for countries that impose a country level high level domain
e.g.
.com.au in Australia or .co.uk for the UK, whereas some
countries do
not do this under their country e.g. .it in Italy.
- Entry/Exit page
- In a session, the entry page is the first URL requested in the
session, and the exit page the last (which is a bit unreliable
as an
indicator).
- Referrer
- IIS can record the link the referred the browser to this web
page. The referring site is the sitename part of the referrer
(i.e. the
page level detail of the referring web page is omitted).
- Broken links
- These are hits that resulted in an error status code. So it
may
be a broken link to the site from within the site, or from
another
site. The broken referrer references indicates the page that
contained
the broken reference. A "-" for the referrer probably indicates
a robot
or a probe searching for a know vulnerability.
- Error
- The following HTTP status codes are not considered to be
errors 200, 201, 202, 203, 204, 205, 206, 304, 301, 302, 303,
307
- Browsers
- This is the "agent string" provided by the browser or robot.
These are then summarized in the "browsers and operating
systems" table
where the agent string is analysed to identify the general
browser or
operating system type.
- Robot
- Robots, a.k.a. spiders or crawlers, are identified in one of 2
ways. The first is by noting the agent string of any browser
that is
fetching the site's "robots.txt" file. The second way is
identify well
known string patterns within the agent string. IISView collects
some
statistics separately for agents classified as not being a
robot, such
as sessions.
- Bounce rate
- This is the percentage of sessions that were less than 3
seconds
compared to the whole.
- Input filter
- For a "site", a filter set controls which IIS log records are
included in the statistics gathering phase. Each record is
tested
against the rules in the filter set and if the field value in
the
filter matches the filter pattern then the record is excluded if
the
rule indicates so.