Show TOC

Optimizing PerformanceLocate this document in the navigation structure


To achieve optimum performance in your system when indexing and classifying documents you should modify the queue parameters in line with your personal usage of TREX. If your usage changes significantly after this first modification, you may have to change the parameters again.

The following settings are relevant for performance.

  • When does the queue server trigger the indexing?

    Parameter: Schedule Type, Schedule Time, Schedule Max Documents

  • How many documents are indexed at one time?

    Parameter: Index Bulk Size

  • When does the queue server trigger optimization?

    Parameter: Optimize Bulk Size, Initial Indexing Mode

  • How many times can a processing step be repeated?

    Parameter: Max Retry Count

To determine the optimum settings, use the following questions to check which scenario applies to you.

  • Do you need to index large data sets initially?

    If this is the case, choose Initial Indexing Mode = On. When the initial indexing run has been completed, set this parameter to Off.

  • Do end users edit larger numbers of documents at larger time intervals? For example, are documents updated weekly?

    If this is the case, you have the following options:

    • Choose Schedule Type = Count and set the Schedule Max Documents parameter to the approximate number of updated documents.

    • Use the Flush function to trigger the indexing of these documents.

    • Set the queue parameters so that indexing only takes place at the weekend.

  • When do you want new or changed documents to be visible?

    If you want them to be visible quickly (for example, within 30 minutes), choose Schedule Type = Time and set the required interval (for example, 30 minutes).

  • Are there times at which the system load is considerably less than at other times?

    If the system is mostly used nationally, the system load will normally be lower outside of normal working times. You should use this time for indexing. For example, you can plan indexing daily at midnight using the Schedule Time parameter.

  • Do you mainly want to index and classify documents with low availability?

    If this is the case, increase the value for the Max Retry Count parameter. This makes sense for indexing external Web sites. Reason: This is because TREX may have to try several times to access the Web pages to be indexed if the Web server in question is overloaded.

    We do not recommend setting the value above 20. If processing has failed that many times, you can assume that the Web site you are trying to index no longer exists.