LogMiner 1.20

LogMiner is a powerful log analysis package for Apache (or other web servers using the combined log format).
LogMiner can extract and present several reports, about visits, hits, traffic, requests, navigation paths, browsers and OSs used by users and so on.
Data is stored in a PostgreSQL database, using a schema which has been optimized to reduce redundancy at minimum.
LogMiner has been inspired by the popular Webalizer package, but it differs in certain core points:
data is stored in a DBMS backend and reports are generated on-the-fly, while Webalizer generates plain html files. A DBMS allows to extract and aggregate data in many ways, whenever you need. A drawback is that you wont have the processing speed of Webalizer when parsing log files.
Webalizer only indexes the last 12 months of data: previous reports are still on disk, but arent accessible from any link. This is the most annoying thing of Webalizer, in my opinion: Im sick of having hundreds of unused files around
LogMiner allows to navigate to previous months easily.
Webalizer reports are hardcoded in the program. LogMiner implements reports in a more extensible way. Each report is in fact a simple PHP class, usually supported by a PL/pgSQL function (although youre free to insert your SQL queries in the PHP code if you like).
LogMiner offers more reports than Webalizer: for instance, the OS charts and the navigation graphs.
Depending on your needs, you might prefer LogMiner over Webalizer, especially if you like having a central SQL repository for your data which enables you to extract the data you need at any time, or to add a kind of report which wasnt planned from the start and apply it to older data.
If you dont have a PostgreSQL database or you expect to process huge log files in seconds, go for Webalizer.
Enhancements:
- This release adds support for libpqxx 2.6.9.
- SQL errors are now logged and dont interrupt the parsing of log files.