Products
Webtrends Analytics 8.x
Webtrends Analytics 9.x
Cause
What are table limits?
The Webtrends analysis and report databases are a hierarchical collection of flat files defining the scope of data processed from log files and displayed in reports. The analysis tables, also known as the FastTrends database, set the upper limit for the total volume of data processed after conforming to global and profile specifications. Log file data that does not meet these specifications, for example, entries intended to be excluded by filters, will be omitted from the tables. The report tables, also known as the On Demand database (ODDB), define the extent by which data from the analysis tables is viewable in Webtrends’ reports in the user interface and as exported reports. The report tables will almost always be a fraction of the analysis tables, no greater than 20% of the analysis table’s limit, though default limits will often be less.
To view resource utilization in existing profiles, edit the profile and navigate to Analysis > Table Sizes. A matrix of reports, their status, and related table information displays:
Analysis Status – indicates if an analysis table is less than 80% full (blue), within 20% of the limit (yellow) or full (red)
Report Status – indicates if a report table is less than 80% full (blue), within 20% of the limit (yellow) or full (red)
Smart Limiting – states whether Smart Limiting, also known as Smart Trimming (or just trimming) is in effect. When a table reaches 90% capacity, the application will prune the groups of oldest, least frequent and least relevant elements (provided that no individual group totals more than 80% of the table contents), to ensure that table space remains available for future analyses. Not all reports use Smart Limiting.
Potential Memory Impact (MB) – The amount of memory used for report generation in the event tables reach capacity. When “Unknown” displays, this indicates no assessment can be made. This value, when known, is useful for identifying reports which may have issues as tables grow in size over time.
Current Memory Impact (MB) – The amount of memory currently used for report generation.
Impact Without Interval Data (MB) – The amount of memory used by the report if interval data is not enabled.
1D Analysis Limit – The maximum number of unique elements tracked for the report’s first dimension.
1D Analysis Count – The current number of unique elements tracked for the report’s first dimension.
1D Report Limit – The maximum number of first-dimensional elements viewable in reports.
1D Report Count – The current number of elements displayed for the report’s first dimension.
2D Analysis Overall Limit – The maximum number of elements in the report, a multiple of the 1D Analysis Limit by 2D Analysis Per 1D Limit values.
2D Analysis Overall Count – The current total number of unique elements analysis table.
2D Analysis Per 1D Limit – The maximum number of second-dimensional elements tracked for each first-dimensional element in the report.
2D Analysis Per 1D Count – The current total number of second-dimensional elements tracked for each first-dimensional element present in the report.
2D Report Per 1D Limit – The maximum number of second-dimensional elements viewable in reports for each first-dimensional element.
2D Report Per 1D Count – The current total number of second-dimensional elements viewable in reports for each first-dimensional element.
How do I know when table limits are a problem?
Table limits that are too low can be identified by reviewing the profile’s table information. In addition to the status indicator displaying as a red ball, the count table will display a value equal to the limit plus one (for first-dimensional tables) or two (for second-dimensional tables). As an example, the 1D Analysis Count may be 10001 or the 2D Report Per 1D Count may be 202. The count beyond the table limits imply there is additional data available that could not fit into the table. When viewing reports, results from the final page of report data drop off from a level exceeding the minimum desired level to be included in reports. An example of this is a Pages report with 2000 entries, and on the twentieth page of one hundred entries each, the 2000th ranked entry has 16 page views. It may be assumed that additional entries will be ranked in descending order and may increase in quantity as popularity decreases, thereby justifying a table limit increase.
Table limits that are too high can cause analysis failure or degrade performance, and this is evident in error messages in the profile’s status log containing error codes such as -1073741819 or -2147483641, which, while not expressly indicative of a problem with table limits, are representative of issues with memory, among which incorrectly configured table limits are one possible cause. One indicator of table limits being set too high is when a profile’s table information displays negative values for the potential memory impact, and these same reports may also display an analysis count in excess of the ones denoting the tables are full. It should be mentioned that these symptoms may be evidence of corruption, but troubleshooting protocol mandates modifications to configuration settings be made first before proceeding to more extreme methods for resolution.
Note: Though the above description applies to both Webtrends On Demand and Webtrends On Premises customers, Webtrends On Demand customers are not able to modify table limits in the user interface.
In either case, for assistance or advice concerning table limit configuration, contact Webtrends Technical Support.