We use webstats at Columbia to gather tons of information from our web server logs, and to produce a variety of reports in HTML format. Some of the key features include a finely tuned input parser that can recognize the strangest log entries from malfunctioning web browsers. IP addresses are resolved into host names using a separate program called logres. During this stage the log records are cleaned up and written back to disk.

Getsats and logres are capable of reading compressed log files to conserve disk space.

Several reports are available from webstats. The Directory Tree, Domain, and Redirect Reports are created in HTML format. The Extract files are simply a set of log records for a particular directory (written and compressed). The Error Reports are sent to individual users via email.

---

Academic Information Systems 212 854.1919 consultant@columbia.edu
last modified on 02/24/04