You can save usage information to log files by running the collect_usage batch or shell script in the control directory of your deployed application on disk. To automate this process, you can set the script to run in Windows Task Scheduler on Windows, or as a cron job on UNIX.

Oracle recommends running usage collection at least every 10 minutes. If your environment is configured to run baseline updates at regular intervals (or otherwise restarts the Dgraph or Assembler frequently) you should configure your logging intervals such that you do not lose more than 10% of the log data stored in memory when a component restarts. For example, in an environment where you run a baseline update on the hour, you should collect usage logs at ten minute intervals starting at 9 minutes past the hour. On UNIX, this might look like the following:

9,19,29,39,49,59 * * * * usr/local/endeca/apps/Discover/control/collect_usage.sh

The log files created by the script are output to the directories configured in the <app dir>\config\script\UsageCollectionConfig.xml file, and are rolled on a monthly basis.