Using SAS Visual Analytics Designer with SAS IT Resource Management Data Sources

SAS Visual Analytics Designer can create reports that use SAS IT Resource Management data. The resulting visualizations complement the reports that were created using the SAS IT Resource Management reporting processes. Specifically, SAS Visual Analytics reports enable you to view interesting relationships between IT measurements and to work with reports that interact with one another.

For illustration, the following examples use data created by the SAS IT Resource Management adapter for VMware vCenter to highlight some of the IT domain intelligence that is possible with SAS Visual Analytics.

This appendix features a scenario where one or more users performed the following tasks:

The CPU Utilization Trends report analyzes the CPU utilization of VMware clusters, hosts, and guests over an 8-day period. A data source filter was used to subset the data from the DetailGuestSysByCluster aggregation table.

The report contains a pie chart, a treemap, two bar charts, a list table, and a time series plot. Each of these reports offers VMware CPU utilization views over designated time periods.

The intelligence that is gathered by viewing the CPU Utilization Trends report for the VMware infrastructure enables you to understand how the CPU utilization of guests contributes to the CPU utilization of hosts and clusters. This insight is essential for VMware capacity planning activities and directly influences how virtual workloads are best allocated to VMware clusters – especially in high-availability VMware environments.

Interactions that are defined between the four charts, the table, and the plot enable filtering to occur. The filter results in dynamic updates to the report objects as you interact with the objects in the report. The following sections discuss the measures for each object and the interactions between them.

When you open the report for viewing, the pie chart and the treemap that are located on the first row of the report show CPU utilization at the cluster level. Each slice in the pie chart represents the utilization for one cluster (for all dates). Each tile in the treemap represents the count of CPUs in each cluster and the CPU utilization for one date (for all clusters). The two bar charts on the second and third rows of the report show CPU utilization for all hosts and for all guests, respectively. The list table and time series plot that are located on the bottom row of the report show the utilization for all resources in the hierarchy.

The following section discusses the interactions that are defined for the report.

Interactions were created between the graphs and the table that allow for filtering to occur as you interact with the graphs. The following two images show the interactions.

The first image shows the defined interactions on the Interactions tab. The objects were manually arranged to show a clear view of the defined interactions.

The second image includes the derived interactions. These images show that the pie chart and the treemap behave like top-level controls.

By selecting a slice of the pie chart, you apply a cluster filter to all of the reports that interact with this chart. Likewise, by selecting a particular date tile, you apply that date filter to all reports that interact with the treemap. Similarly, selecting a bar at the host level applies a host filter to the guest bar chart as well as to the list table and time series plot that are at the bottom of the report. Selecting a bar at the guest level applies that filter to the list table and the time series plot.

Click on a slice in the pie chart to update the other report objects dynamically. For example, the following display shows that the blue slice from the pie chart is selected. Because the pie chart selection corresponds to VIRTUAL DESKTOPS 2, the other report objects are dynamically updated with the incoming filter. The Properties tab for Bar Chart for Guest (the yellow bar chart) shows that the filter VIRTUAL DESKTOPS 2 has been applied.

Similarly, selecting a particular date from the treemap further filters all of the report objects except the pie chart. This same downstream behavior occurs for the host- and guest-level bar charts as well as the crosstab report. On the Properties tab for the time series plot, the following selections were made:

The Utilization and Datastore Capacity report analyzes the data storage capacity for clusters and the CPU, memory, disk, and network usage for guests for a single day at a time.

The report contains a pie chart, a bar chart, a treemap, and two list tables. The pie chart shows the number of guests in the selected cluster, and the bar chart shows the number of guests in the selected host. Selections in the pie chart or the bar chart as well as the drop-down list control filter for the Date field are used to filter the measurements of interest.

The intelligence that is gathered by viewing the Utilization and Datastore Capacity report consists of these items:

This report enables VMware administrators to assess the day-to-day performance of the VMware infrastructure. Using this report, VMware administrators can quickly identify over- or underutilized guest systems based on their computing performance and capacity.

The following image shows the defined and derived interactions of the Utilization and Datastore Capacity report:

The following image shows the Properties tab for the treemap. The Incoming Filters list shows that the following filters were selected:

Filters are applied to the tiles in the treemap and to the rows in the two list tables.

The following image also shows a data tip on the treemap that contains the name of the selected guest and additional measures that were added as data tip values.

While interacting with reports, you can enlarge any report object to get a clearer view of the measurements. For example, you might want to expand the treemap that contains all guests within the cluster before any filters are applied.

The CPU Utilization for the Latest Date report analyzes the CPU utilization of VMware clusters, hosts, and guests over a 24-hour period. In this example, a data source filter was used to subset the data from the DetailGuestSysByCluster aggregation table. For more information, see Data Source Filter.

The report contains two sections. The first section of the report shows a bar chart, time series plot, and crosstab table. The second section of the report contains two dual-axis time series plots that show related measurements. The CPU and memory usage percent and the disk and network usage rates are measurements that are most relevant at the guest level.

The intelligence that is gathered by viewing the VMware Cluster CPU utilization in the CPU Utilization section is the amount of CPU that is used by each cluster, host, or guest (observed at each level of the hierarchy). The value for CpuUsedMSec is available as the height of each bar and the position of each line and as a column in the crosstab table.

The intelligence that is gathered by viewing the VMware Cluster CPU Utilization in the Guest Measurements section is the CPU busy and memory usage percent measurements and the disk and network usage rates for the guest that was selected.

Interactions within the report and between report sections enable you to drill down from cluster to host to guest. You can also link to another section of the report that shows additional measurements at the guest level. The following image shows the interactions between the report objects in this section as well as the link to the Guest Measurements section.

If you click any bar in the bar chart, the brush interaction highlights the measurements for the corresponding bar in the crosstab. This image shows that the cluster VIRTUAL DESKTOPS 2 is highlighted.

If you double-click on a bar, you can drill down into the bar chart. Similarly, you can double-click on a line to drill into the time series plot. The following image shows the hosts for the VIRTUAL DESKTOPS 2 cluster.

You can continue to drill down into the guest level. The following image shows that the bar for the VIRTDESK14 host is selected and that the bar for the VMWWIN8189 guest is highlighted. It also shows that you can interact with a bar and select the action to move to another section within the report.

When you double-click on a bar for a guest and then select the section link, the report is redirected to the Guest Measurements section with the incoming filter applied. This section contains two dual-axis time series plots. The first plot shows the CPU busy and memory usage percent measurements for the guest that was selected. The second dual-axis time series plot shows the corresponding disk and network usage rates. The Properties tab shows the filters that were applied from the section link.