dcsimg

Best practices for FTP data sources

Products

Webtrends Analytics 8.x
Webtrends Analytics 9.x

Cause

When Webtrends Analytics analyzes log files from an FTP data source, it does not maintain the FTP cache between analysis cycles. Which means that every profile on every analyzes will download the logs the data source lists. If the FTP location holds a large amount of data. This can impact analysis performance.

Resolution

When using an FTP data source on a profile, consider the following guidelines.

Date macros:
With date macros being placed within a data source, only the most recent logs will be downloaded for analysis. As a safety precaution, it is advised that the date macros be set to retrieve the last four days of data. This will prevent any interruptions in analysis from creating gaps in the reported data.

Localization of log files for long term reanalysis:
In the event that a reanalysis for a large time frame is needed, date macros are not an option. For example, if the last two years of historical data need to be reanalyzed or analyzed again in a new profile, all of those logs will need to be brought to the machine. To keep the data source from having to re-download the files between analysis cycles. It is best to temporarily bring the log files to a local drive on the machine and change the data source to use the local path. If the log files in question are too large for the analysis machine, they can be brought over in sections (in the example above, perhaps a month at a time) and then removed for the next section.

More Information

For further information on date macros, refer to KB 5000001819How do I set up a data source in Webtrends Analytics Software?