Response Code

Hello,

I've got following information for my web site from the Webalizer
statistics. I wonder if there is any clue in the data about what need
to be done for improving the pages, or if there is anything wrong with
the site.

Hits by Response Code
Code 200 - OK 12033
Code 206 - Partial Content 615
Code 301 - Moved Permanently 21
Code 304 - Not Modified 2362
Code 400 - Bad Request 1
Code 401 - Unauthorized 56
Code 403 - Forbidden 1440
Code 404 - Not Found 1894
Code 405 - Method Not Allowed 3

My web site is not a commerical one. It contains only discussions on
Chinese language planning. It does get many hits:
www.pinyinology.com

Thanks in advance for expertise.

dungping

dungping wrote:

Hello,

I've got following information for my web site from the Webalizer
statistics. I wonder if there is any clue in the data about what need
to be done for improving the pages, or if there is anything wrong with
the site.

Hits by Response Code
Code 200 - OK 12033
Code 206 - Partial Content 615
Code 301 - Moved Permanently 21
Code 304 - Not Modified 2362
Code 400 - Bad Request 1
Code 401 - Unauthorized 56
Code 403 - Forbidden 1440
Code 404 - Not Found 1894
Code 405 - Method Not Allowed 3

My web site is not a commerical one. It contains only discussions on
Chinese language planning. It does get many hits:
www.pinyinology.com

Looks pretty normal to me. But without knowing which documents generated
which codes, there's not much more to say.

A few hints:

'More 304 is good'. It means that unmodified documents are not repeatedly
served to the same client or are cached by a proxy, thus reducing the load
on your server. If you are using static HTML documents, the server should
take care of this. When using dynamically generated documents, you may have
a look at the code, if it generated 'Last-Modified' and/or 'ETag' response
headers and recognizes and handles 'If-Modified-Since' and/or 'If-Match'
request headers.

You should have a closer look at which documents generated 4xx codes. Many
4xx hits are generated by script kiddies testing for vulnerable software on
your server that is not present or vulnerabilities of the server itself.
But it could also indicate broken links on your site.

--
Benjamin Niemann
Email: pink at odahoda dot de
WWW: http://pink.odahoda.de/

On Sat, 1 Jul 2006, Benjamin Niemann wrote:

Hits by Response Code
Code 200 - OK 12033
[...]
Code 304 - Not Modified 2362

Looks pretty normal to me. But without knowing which documents
generated which codes, there's not much more to say.

A few hints:

'More 304 is good'. It means that unmodified documents are not
repeatedly served to the same client or are cached by a proxy, thus
reducing the load on your server. If you are using static HTML
documents, the server should take care of this. When using
dynamically generated documents, you may have a look at the code, if
it generated 'Last-Modified' and/or 'ETag' response headers and
recognizes and handles 'If-Modified-Since' and/or 'If-Match' request
headers.

Indeed (it's "If-None-Match", to be accurate).

I'd recommend reading Mark Nottingham's cacheing tutorial
http://www.mnot.net/cache_docs/
and use the cited "cacheability engine" to report on the site,
to help identify any areas of concern.

You should have a closer look at which documents generated 4xx
codes. Many 4xx hits are generated by script kiddies testing for
vulnerable software on your server that is not present or
vulnerabilities of the server itself.

Yes, we see lots of those. But also garbled URLs that appear to be
defective web searches. This makes it quite hard to spot genuine
problems.

But it could also indicate broken links on your site.

Yes, but IMHO those can be more accurately identified by running a
link checker on your site.[1]

4xx errors in the log can also indicate broken links to your pages
from *somebody else's* site. In which case you could alert the other
webmaster, and/or set up a redirection to the correct URL on your own
site.

h t h

[1] If you run a windows system, then Xenu link checker is amazingly
fast, and generally to be recommended. However, it uses Windows
routines "under the covers" to access URLs, and there's a couple of
kinds of URL error which Windows silently fixes-up "under the covers",
with the consequence that Xenu checker can't see the error. For
example, using "\" instead of "/" in a URL path. So it's recommended,
but not ideal.