Optimizing Microsoft SQL Server Analysis Services: Optimization Tools: Basic Usage

As we have discovered, we are offered an opportunity to set criteria for filtering the results that are returned for query usage analysis in the report we have specified in the previous dialog. By setting filter criteria, we are able to place restrictions on the data that is returned, via the Usage Analysis Wizard, from the query log, which is the source of the cube data we are analyzing (we will discuss the log, and mention other options for its use, in later articles). Available filter criteria, and a brief description of what each entails, are included in Table 2.

Table 2: Filter Criteria Available in the Usage Analysis Wizard

As we noted earlier, this step is optional. The report we selected in the previous dialog will be returned with no filters if we simply leave the criteria boxes in the dialog unchecked / uncompleted, and proceed by clicking Next.

5. Select the Queries That Ran More Than option by clicking the checkbox to its left.
    
6. Select 6 in the “times” selector box.

We have indicated that we wish to limit the Query by Hour Graph report to include only those queries that have been executed more than six times, as depicted in Figure 8.

Figure 8: The Select the Criteria to Filter This Report Dialog with Our Selection

7. Click Next.

The viewer appears, displaying the Query by Hour Graph report, as shown in Figure 9.

Figure 9: The Filtered Query by Hour Graph Report in the Viewer

The Query by Hour Graph report displays, after filtering out queries that have been executed 6 times or less, the total number of queries processed. The queries are grouped and displayed by the respective hours of day at which they were processed. Even though the example display in Figure 9 displays minimal data, we can still get a feel for how the results are presented. (Keep in mind that your results may differ dramatically, depending on the history of the cube upon which you run the Usage Analysis Wizard.)

My intent in using the sample cubes, in many of the exercises we will transit throughout the series, is to ensure that any reader with an MSAS installation might readily complete the practice steps. The samples present limitations, obviously, from the dual perspectives of “real world” statistics and imagination in general. Be sure to put the steps we discover together to work within the background of a robust implementation, within your organization or elsewhere. For example, in our current case, get a copy of a large cube log and run the various report types for some excellent hands-on practice.