Greatest Deviation Widget

This widget shows Jobs with a high run-time variation over the selected time frame to find unexpected behavior or trends.

This widget helps you identify tasks with a high deviation in one of the numeric attributes. You can also use the drill-down capabilities to display the individual run-time evolution and other details.

The values that are displayed to the left of the axis represent the deviation of failed jobs. The values that are displayed to the right represent the deviation of successful jobs.

Screenshot that shows a New deviation chart which is a Greatest Deviation Widget. This widget displays jobs with a high run-time variation to identify unexpected behavior or trends.

  • Appearance

    • Title: set the title for the widget (displayed in the top bar of the widget)

      If no title is defined, the default title Greatest Deviation is used.

  • X-axis: here you select the values to be displayed on the horizontal axis. You can select the following values:

    • Show: select how you want to categorize the data (by agent type, hour of day, status code, and so on.)

      Note: The deviation cannot be calculated for clients.

    • Select the Filter categories by deviation factor checkbox if you want to show only values from a certain deviation percentage onwards.
      • Enter an absolute deviation percentage in the picker (example, 150 percent).
      • Select if you want to include positive deviation values, negative ones or both.
  • Y-axis: here you select the values to be displayed on the vertical axis.
    1. First, select how you want to show the values, by runtime (real runtime) or by ERT (estimated runtime).
    2. Select which kind of jobs you want to be analyzed: ended or started.
  • Advanced configuration

    • Here you can set the sampling coefficient of variation to calculate the deviation. This coefficient allows you to compare data sets with different means.

      For example, by setting the coefficient to 80 percent, you would compare the average (mean) run time/ERT of the last 20 percent of executions that are filtered by a sub-component (example object name) with the average run time/ERT of the first 80 percent.

      Note: The percentage is taken from the number of executions.

      Tip: The recommended setting for the coefficient is 80 percent (default value).

Note: Hover over the bars to see the exact number of reported elements.

Note: You can remove one or more bars by clicking them, selecting the exclude selected elements checkbox and clicking Apply.

  • Description: optionally, you can specify a description to better identify the data that is displayed.

Setting Widget Specific Timeframe

You can set a specific local timeframe directly in a widget. When this feature is activated, all data contained in the widget is specific to your set timeframe. All other widgets in your dashboard are unaffected by individual widget changes.