The Apache log-file format is pretty standard and there are CPAN tools which can easily parse and analyze it.

A Perl script could be launched to run at regular intervals, e.g. by a crontab in Unix, which would scan the current log-file and build up answers concerning it. This script would run in the context of a local user, not the Apache-server userid, and thus would be able to see and interpret the contents of the log file. (It could, of course, place its findings into a file that a web-page could “see.”)

Again, “you are not the first soul to have thought of this,” so search very-aggressively through CPAN for tools that can help you. In many ways the greatest strength of this language is not the language itself, but rather CPAN.