Metric Extrapolation View for Groups

Available only in vFabric Hyperic

Extrapolating Metrics for a Group

This page allows you to predict future values of a metric for the members of group based on historical metrics. 

The Metric Data Extrapolation plugin is available on the Views tab when a group is selected. Although this function is available for any group, whether compatible or mixed, metric projections are most meaningful for compatible groups.

The extrapolation method is a simple linear regression using the Apache Commons Math package. The historical metric data for a group is used to calculate a slope, which is used to predict future values. For more information on the simple linear regression algorithm, see http://commons.apache.org/math/userguide/stat.html.

To extrapolate a metric for a members of a group: *

  1. Navigate to compatible group of interest.

  2. Choose the Metric Data Extrapolation option from the Views tab.

  3. You can select:

    • Metric to extrapolate.  By default, Availability is selected. If desired, choose another metric.

    • Known Data Range. The amountof history upon which the extrapolation will be based. You can select values ranging from 1 day to 1 year. Typically, a longer data range yields a better extrapolation. In some cases, however, it might be better to select the data range based on the timing of environment-related events, such as a new release or infrastructure change.

    • Projection Range. The total duration included in the trend analysis, including the Known Data Range and the Projection Range. The range you select determines, but is not equal to, how far into the future the metric will be projected.

    • Threshold value for the metric. Charts will display a horizontal line at the metric value you enter.

  4. Click Update to plot the extrapolated values for the selected metric for each member of the Group. Historical values are plotted in blue.

Understanding Extrapolation Charts

Each chart has the same scale, to better show differences between the resources, and ease analysis. This is different from other HQ charts which are scaled independently. The heading for each chart shows:

  • The resource name and metric charted

  • The units of the metric charted

  • Extrapolation confidence; values can be 'Excellent', 'Good', 'Average', 'Questionable', or 'Poor'.

Extrapolated values are represented by the orange triangle that appears between "now" and the end of the projection range on the timeline.