Simon Redding <[EMAIL PROTECTED]> wrote: > On the other hand, I'm trying to establish concurrency for a pre- > packaged J2EE web app (to record timesheets) which has Apache as a > presentation tier. All the users are on a flat corporate Intranet, > which has no proxies, IP address masking etc etc that would skew > session/visit information. Normal time to complete the application > (i.e. fill in a timesheet) would be about 10 minutes, with quite a > lot of page requests in that time (it's not a great app). All I want > to know is the number of 'concurrent users', which I would take to > mean the number of sessions that have not timed out on the keep-alive > (or at the application level are from an IP address which has sent a > request in the last (say) 10 minutes). > > Can analog do this?
No. > Can anyone recommend a better package/solution? The place to count "concurrent sessions" is in the application itself. but if you have to do it from the logs, it's not that difficult if there is a "log out" page that people use. It shouldn't be that difficult to write a script that goes through the logs and increments a counter when it sees an entry for the "log in" page and decrements it when it sees an entry for the "log out" page. But it won't be able to tell when a user doesn't bother logging out, and just times out. If you log a "sessionID", you could keep a running count of "live" sessionIDs, which is a bit more complicated - you just need to keep a list of live sessionIDs, along with the last time they had an entry in the log, and then discard them as the time stamp exceeds the session timeout. (It's not hard to program, as long as you don't have to deal with millions of concurrent hits, it's just a pain to get it right). Again, the best way to do this is to have the application itself create a separate log, so that when a session times out, rather than the user logging out, the application can still create a "logged out" entry. Trying to drag this kind of information out of the logs after the fact, when it wasn't explicitly put into the logs in the first place, is more trouble than it's worth. Aengus +------------------------------------------------------------------------ | This is the analog-help mailing list. To unsubscribe from this | mailing list, go to | http://lists.isite.net/listgate/analog-help/unsubscribe.html | | List archives are available at | http://www.mail-archive.com/[email protected]/ | http://lists.isite.net/listgate/analog-help/archives/ | http://www.tallylist.com/archives/index.cfm/mlist.7 +------------------------------------------------------------------------
