Setting up performance cubes for Extended Reporting

A performance cube is a multidimensional structure into which data is compiled for Extended Reporting. Default performance cubes are installed with the BMC Real End User Monitoring Experience Aggregation Server for Extended Reporting component. The following topics describe how to define additional performance cubes and specify the retention of that data.

Performance cubes

Performance cubes are defined according to three aggregation subjects.

• Applications
• Infrastructure
• Users

Normally, we think of cubes as having three dimensions. However, for the purposes of multidimensional analysis, a cube can have as few as one and still provide useful results. Cubes also have a time dimension and a data-provider dimension automatically.

Performance cube aggregation subjects

You can set up to three performance cubes per aggregation subject (in other words, a maximum of nine cubes).

• Application cubes—Tracking every page hit to the application, the application aggregation cube is the basis for understanding the relative health of pages, sites, or specific parameters, such as search type. For example, it might contain performance metrics for every combination of site name, page name, and region every five minutes. Application cubes use page transcripts as source data.
• Infrastructure cubes—Tracking every page hit to the equipment and platforms on which a site runs, the infrastructure cube allows operators to identify observed capacity, chronic problems with tiers of infrastructure, and understand how various platforms work together. For example, it might store performance metrics for every combination of load-balancer virtual IP address, Web server ID, and application server. Infrastructure cubes use page transcripts as source data.
• User cubes—Tracking every visit from every user, the user cube is the basis for customer service level reporting and activity by region. For example, it might contain performance metrics for every combination of company, user name, and geographic region in a particular day. User cubes use session transcripts as source data.

Default performance cubes

The following figure illustrates the default performance cubes installed with the Aggregation Server component.

Performance cube filtering

Filtering allows you to exclude data from aggregation. For example, you might want to exclude QA data (such as load tests). You can define filter expressions to compare or match data against specific values or to check for the existence of specific values in the data. For details about filter expressions, see Using-the-Expression-Builder-to-create-a-filter.

Before you begin

1. Log on to the Aggregation Server for Extended Reporting web interface.
2. Click Performance cubes - setup in the Data flow settings.

To create or edit a performance cube

1. Perform one of the following actions:
   • In one of the three aggregation subject sections, on the Action menu, click Add.
   • On the shortcut menu for a given performance cube, click Edit.

2. Type a cube name. Choose a meaningful name because the name will appear throughout the Aggregation Server for Extended Reporting user interface and can appear in reports.

   Success

   Tip

   Assign cube names that reflect their component dimensions. For example, website-page-geo denotes website, page, and geographical dimensions.

3. To prevent the cube from growing to a point where performance and retention are negatively affected, leave cardinality management enabled for the cube. If you want to disable cardinality management, clear the Use Cardinality Management check box.

4. If you want to filter transcript data prior to aggregation, type a filter expression in the Filter expression box. For help with composing a filter expression, click Build an expression to use the Expression Builder tool. For related information about the Expression Builder, see Creating-a-filter.

   Success

   Tips

   • To reduce cardinality, use "is_container" to avoid documents, which is useful only if an operator wants to track a series of documents.
   • You can also configure the Analyzer to filter out documents.
   • The operators that you can use to compare, match, or search depend on the data type of the field. You can use logic keywords, such as AND, OR, and NOT, to combine filter expressions, and use separators, such as parentheses and brackets, to group parts of the expression.
5. For each dimension, click Add and perform the following:
   1. In the Dimension alias box, type a meaningful name for the dimension.
   2. In the Dimension source box, type the field identifier.

   3. To select a field identifier from a list of observed fields, click the lookup button and then click the field identifier.

      Success

      Tip

      Avoid adding fields such as "idle time," which will explode the size of the cube.

   4. To store the data by using the original case of the data, select the Case sensitive check box. To store the data in lowercase, do not select the Case sensitive check box.

      Success

      Tip

      Select Case sensitive only if there is a valid requirement.

6. Click OK.

To enable or disable a performance cube

In the row for a given cube, click ON or OFF.

To delete a performance cube

1. In the row for a given cube, in the shortcut menu, click Delete.
2. Perform one of the following actions:
   
   • Click Purge all for this cube.
   • Click Do not purge data for this cube.
3. Click OK.

To purge data from a performance cube

1.  In the row for a given performance cube, in the shortcut menu, click Purge.
2. Perform one of the following actions:
   • Click Purge all for this Performance Cube.
   • Click Purge data up to and excluding, and select the year, month, and day.
3. Click OK.

Related topics

Retaining-performance-data-for-Extended-Reporting

Monitoring-database-status