Bash Webstat for Apache Logfiles

I am sure you know the annoying “accept cookies from this site” popups. In deed all other availabel webstat solutions require cookies as well as additional code that needs to be inserted into your webpage. This solution is different. You only need your Apache logs, something that should be available for every website running the predominant and most widespread web server on the web, Apache, usually via exclusive ftp download. Though we at elstel.org do also have an online python-CGI/mysql GUI solution for analyzing Apache logs, this solution is most simple and basic. However it requires downloading the logs first. Nonetheless this can also be seen as advantage as it inclusively leads to a log backup for several years possibly spanning multiple server installations with different log formats.

The solution we are talking about is a simple bash command line script to view the visits of the last day. Page views can be listed by URL or by country. Search engine results are grouped by URL and clustered with a count. If you have a certain, possibly wildcarded URL, you can view the countries people came from visiting this URL as well as the whole timeline of the visit, the browser in use, the country and of course the pages that have been visited in sequence. Visits can not only be viewed for a contained URL but also for a given country. Downloads and search engine crawls can be viewed separately. Most times this information is viewed for a special day given by date although an aggregation feature for the most recent log files is also available. Just have a look at the command lines below.

Download:
webstat-bash v1.4.2 new llogrotate accomplished (formerly llogrotate-ng), interesting rationale about program verification in README
webstat-bash v1.4 python3-compatible (Windows), nameclash itemcount/count fixed (Linux), faster visitor table
webstat-bash v1.3 logview for multiple domains, logrotate without parameters, upd. search engine & download recognition
webstat-bash v1.2 --at back-N: stats for more than last day, visits: also view accesses with status code 4xx and others
webstat-bash v1.1 better spider and bot recognition; differ between Google-search and Gg-Ads hits
webstat-bash v1.0 first published version
software/SHA512SUMS.signed.
Please sign our Contributor License Agreement if you want to contribute code. Otherwise we can not assimilate and re-distribute your changes here at elstel.org



Query Examples

> alc --at 21/Feb/2014
…
GeoIP Country Edition: AT, Austria  4
GeoIP Country Edition: IN, India  4
GeoIP Country Edition: RU, Russian Federation  5
GeoIP Country Edition: JP, Japan  6
GeoIP Country Edition: FR, France  6
GeoIP Country Edition: UA, Ukraine  7
GeoIP Country Edition: CA, Canada  8
GeoIP Country Edition: CN, China  9
GeoIP Country Edition: DE, Germany  31
GeoIP Country Edition: US, United States  33
sum: 153

A number of 153 different IPs have visited on that day with different numbers of IPs belonging to different countries.

> alc --at 21/Feb/2014 pages
…
8 /qemu/
10 /html5video/Html5VideoScripting.html.en
10 /xchroot/
11 /index.html.en
11 /html5video/FlashVersusHtml5Video.html.en
14 /FilmReviewSamsara.html.de
55 /OS2Warp/InstallUpdate.html
sum: 175
pages 172/3 - pages/rss

172+3=175; ‘rss’ amounts to ViewRSS.php and elstel.rss. These are page accesses rather than IPs. A page can be retrieved multiple times from the same IP.

> alc --at 21/Feb/2014 seo-landing
…
4 google				/html5video/Html5VideoScripting.html.en 
6 google				/html5video/FlashVersusHtml5Video.html.en 
6 google				/qemu/ 
7 google				/xchroot/ 
12 http://de.wikipedia.org/wiki/Samsara_		/FilmReviewSamsara.html.de 
13 google				/OS2Warp/InstallUpdate.html 
sum: 100
> alc --at 21/Feb/2014 referers
…
3 http://darkbooks.org/  
12 http://de.wikipedia.org/wiki/Samsara_(2011)  
44 google 
sum: 100

Result counts by search engine and referer.

> alc --at 21/Feb/2014 --page elstel.rss
GeoIP Country Edition: DE, Germany  1
GeoIP Country Edition: AT, Austria  1
GeoIP Country Edition: EU, Europe  2
sum: 4

Now if you are interested in different IPs/visitors visiting a site rather than retrieval access counts this command gives you what you want, grouped by country.

> alc --at 21/Feb/2014 --page xchroot visits
…
192.100.130.7 (GeoIP Country Edition: EU, Europe) Mozilla/5.0_(Windows_NT_6.1;_WOW64;_Trident/7.0;_rv:11.0)_like_Gecko::
0 21/Feb/2014:18:36:36 /xchroot/ 			 200 <<://www.bing.com/search?q=x11+in+chroot&go=&qs=ds&form=QBRE
1 21/Feb/2014:18:36:36 /elstel.rss 			 200 << http://www.elstel.org/xchroot/

> alc --at 21/Feb/2014 --from EU visits
192.100.130.7 (GeoIP Country Edition: EU, Europe) Mozilla/5.0_(Windows_NT_6.1;_WOW64;_Trident/7.0;_rv:11.0)_like_Gecko::
0 21/Feb/2014:18:36:36 /xchroot/ 			 200 << http://www.bing.com/search?q=x11+in+chroot&go=&qs=ds&form=QBRE
1 21/Feb/2014:18:36:36 /elstel.rss 			 200 << http://www.elstel.org/xchroot/

145.62.32.131 (GeoIP Country Edition: EU, Europe) Mozilla/5.0_(Windows_NT_6.1;_WOW64)_AppleWebKit/537.36_(KHTML,_like_Gecko)_Chrome/30.0.1599.101_Safari/537.36::
0 21/Feb/2014:07:39:04 /OS2Warp/InstallUpdate.html 			 200 << http://de.wikipedia.org/wiki/OS/2
1 21/Feb/2014:07:39:04 /elstel.rss 			 200 << http://www.elstel.org/OS2Warp/InstallUpdate.html

Finally this lists up visits for certain criteria. A visit are all page recalls from the same IP and browser on that day. Different browsers (agents) from the same IP make up different visits.

> alc --at 06/Jan/2020 downloads
/database/dbschemacmd-v1.1.tar.gz				Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/79.0.3945.88 Safari/537.36 200
/database/dbschemacmd-v1.1.tar.gz				Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/79.0.3945.88 Safari/537.36 200
/confinedrv-v1.7.7				Mozilla/5.0(Linux;Android 5.1.1;OPPO A33 Build/LMY47V;wv) AppleWebKit/537.36(KHTML,link Gecko) Version/4.0 Chrome/42.0.2311.138 Mobile Safari/537.36 Mb2345Browser/9.0 301

Get stats for the latest days present in the last 3 logfiles:

> alc --at back-3 first-day
earliest date: 02/May/2021 00:00
02/May/2021:00:00:38

> alc --at back-3 --page /atea/ 
earliest date: 02/May/2021 00:00
GeoIP Country Edition: FI, Finland  1
GeoIP Country Edition: AT, Austria  1
GeoIP Country Edition: US, United States  2
sum: 4

Viewing a whole visit for visits with more than 17 pages / setting the visit view limit:

> alc --at back-3 --limit 60 visits 34.77.160.243
earliest date: 02/May/2021 00:00
34.77.160.243::
0 16/May/2021:02:17:05 / 			 301 << -
1 16/May/2021:02:17:05 / 			 301 << -
2 16/May/2021:02:17:05 / 			 301 << -
3 16/May/2021:02:17:05 /index.html.en 		 200 << -
4 16/May/2021:02:17:05 / 			 301 << -
5 16/May/2021:02:17:05 /Contact.html 		 200 << -
   …

There are some more commands for alc which can be shown by alc --help.




post-Installation instructions

Before you can use our solution for your own log files you need to specify the log format. For each location (in this example: ’/home/weblog/[dotplex/|/revido/old]’) you specify the praefix of the logfile names (‘access.log’ or ‘access_log_’), their sort order, their format and possibly an url-match token if access for multiple sites end up in the same logfiles:

#format: M-domain, I-IP, D-date, C-command line with GET/HEAD, S-http-status (200=ok), L-content/file-length, R-http-referer (site the user visited before this one), A-agent(browserid) ‘-’skipped
logdirs=("/home/weblog/dotplex/ access.log .+2. MI--DCSLRA https?://(www.)?elstel.(org|com)" "/home/weblog/revido/old/ access_log_ _-2_w-1---1. I--DCSLRAM https?://(www.)?elstel.(org|com)");

I will give you an example for an ‘MI--DCSLRA’ log entry:

elstel.com:443 77.75.78.163 - - [03/Jan/2021:02:59:55 +0100] "GET /robots.txt HTTP/1.1" 200 3614 "-" "Mozilla/5.0 (compatible; SeznamBot/3.2; +http://napoveda.seznam.cz/en/seznambot-intro/)"

… and here an example for a ‘I--DCSLRAM’ log entry:

195.191.196.90 - - [09/Oct/2012:02:29:04 +0200] "GET / HTTP/1.1" 200 1361 "-" "check_http/v1.4.15 (nagios-plugins 1.4.15)" www.elstel.com

The sort order field for ‘access.log.NN[.gz]’ is ‘.+2.’: This means take the field starting with the second dot ‘.+2’, sort ascending ‘+’ and read until another dot is encountered. The + effectively means to read access.log.1 for more recent logs before access.log.2.gz. If no tag is found it is the most recent logfile, i.e. ‘access.log’.

We also have a more difficult example for logfile ordering:

access_log_2012_w53-0.gz
access_log_2013_w10-0.gz
access_log_2013_w11-0.gz
access_log_2013_w12-0.gz

Here we need to read until the second ‘_’ and take the field until another ‘_’ in descending order in order to start with the most recent log (‘_-2_’). Besides the year there is another field for the month starting with the first ‘w’ and ending with ‘-’ (‘w-1-’). The last field seems to be always zero here but it will also be considered for the sort order by taking the first ‘-’ until the next ’.’ (‘--1.’), all in descending order ‘-’.

Adjust the line with logdirs directly in the sources of the alc script with an editor.

Finally you need to be in the right logfile directory as specified in the logdirs-line when you invoke alc.