Would like to collect and display usage statistics. Who is using system and when. This would be to justify use of system to management. Would like to show numbers as to how many are actually using the system.

5/7/09 Here is an example of a script used to create statistics information on confluence users:
There are no access_logs for confluence. However searching the apache access_log on the proxy shows following for september:
grep '\/confluence\/login\.action' fchssl-access_log.1|wc -l
4633

There may be a combination of GET and POST with login for Confluence.
Thus one would have to cut count about in half.
grep 'GET' temp.txt|wc -l
2662

For jira, it appears that we could get login stats - ie. how many login - by counting number of calles to login.jsp in the access log for tomcat.
grep '\/jira\/login\.jsp' access_log.2008-05-*|wc -l
4564

Ideally we could get statistics from a table with a sql statement such as :
select count(distinct(username)), to_char(calltime,'Mon YYYY')
from apache_access
where calltime>to_date('1-01-2009','MM-DD-YYYY')
and responsecode not in (302,404)
group by to_char(calltime,'Mon YYYY')

So, we'd be recording call-time, username, response code (http response code for call), etc. If there were other types of calls that the site allows, such as calls to perl or php routines or python, etc., you could have a 'url' column, and add to the query a select on the type of ur.
Example:
...
and url like '%.php%'
...
Sample Output:
587 Apr 2009
1480 Feb 2009
1500 Jan 2009
1459 Mar 2009

and responsecode not in (302,404)