Decisions dashboard

 New in 18.0.2  The Decisions dashboard present tables and charts which display various statistics about your decisions.

Important: Do not delete the predefined dashboards and their underlying widgets and searches, otherwise your dashboards will not work any more.

Opening dashboards and widgets

  1. To open Kibana, from the IBM® Cloud Private administration console, click Menu > Workloads > Helm releases.
  2. Search for the name that you entered for your release, for example bai.
  3. In the Helm release page, find the Kibana URL.
  4. To open a dashboard, in the Kibana window, click Dashboards, filter on meaningful words for the dashboard or widget that you want to display, and select it from the resulting list.
  5. To open a widget, in the Kibana window, click Visualize, filter on meaningful words for the widget that you want to display, and select it from the resulting list.

The predefined Decisions dashboard displays technical data from the Operational Decision Manager event emitter about your decisions. It does not show your business data, such as the input and output of your decisions. To learn how to populate this technical dashboard and create dashboards that reflect your business data, see IBM Operational Decision Manager samples. Use the Kibana selection facility on charts or in tables to filter out data. You can later remove the filter by using the Kibana user interface at the top of the dashboard. To select a time range, use the time range picker " ".

The Decisions dashboard displays several widgets. To find find them, search the Visualization page of the Kibana window on the Decisions - prefix. Use the widgets as a starting point to create your own dashboards.

Each ruleset has a dedicated color in the various charts. You can apply a different color to a selected ruleset. You can also filter results on any ruleset by hovering on its name and clicking the Filter this value buttons Filter this value.

General metrics

Help
This widget gives a brief description of the dashboard and points you to IBM documentation for more information. To read it, open it up full screen, then minimize it again when you are done.
Ruleset selector
From this widget, you filter the rulesets for which you want to see the resulting decisions. From the menu, select the ruleset names, then click outside the menu.

Click Apply changes for the filter to take effect and display the corresponding decisions in the dashboard. To disable the filter, click Clear form or use the Kibana user interface.

Service health
This widget displays a pie chart that shows how the decisions are distributed by type of execution results.
  • The EXECUTION_SUCCESS part counts all the decisions that ran successfully and provided the expected event.
  • The TRACE_GENERATION_ERROR part counts the decisions that ran successfully but the associated trace failed to be generated.
  • The EXECUTION_FAILURE part counts the decisions that failed to complete.
To learn more about the execution and trace generation errors, scroll down the dashboard to view the error widgets and read Errors.
Rate
This widget gives the number of decisions per second. This is an average count which is computed from the selected time range. This widget is correlated to the Timeline widget which shows the distribution in time of the same decisions.
Count
This widget gives the total number of decisions in the selected time range.
Timeline
This widget displays a bar graph in which each bar shows how many decisions per second were taken during the period that is marked by the displayed time stamp. The graph shows a maximum of 10 rulesets. Therefore, it does not reflect the exact detail of the rate number that applies to all rulesets.

When you hover over a bar in the graph, a tooltip displays the number of decisions per second, the ruleset, the time stamp, and the bucket size for the time unit considered. The bucket size represents the number of decisions that were grouped to be represented as the same bar. This size depends on the time range filter level. You can zoom in on a bar by selecting an area around it. The represented period and the time range picker " " reflect the change accordingly.

Ruleset execution duration
This widget displays a bar graph in which each bar shows the average duration of ruleset execution for the period that is marked by the displayed time stamp. It shows a maximum of 10 rulesets. When you hover over a bar in the graph, a tooltip displays the average execution duration, the ruleset, the time stamp, and the bucket size for the time unit considered. This size depends on the zoom level.
Note: This average duration reflects only the execution of the ruleset. It does not include the duration of the event emission process.
Tasks table
The table columns reflect the following data.
Tasks
The task name, in the following format: RuleflowName>TaskName.
Rulesets
The ruleset name, in the following format: RuleAppName/RuleAppVersion/RulesetName/RulesetVersion
Executions
The number of executions of this task.
Rules table
The table columns reflect the following data.
Rules
The rule name. For decision tables, the rule name is suffixed with _rowNumber.
Rulesets
The ruleset name, in the following format: RuleAppName/RuleAppVersion/RulesetName/RulesetVersion
Executions
The number of executions of this rule.
Rulesets table
The table columns reflect the following data.
Rulesets
The ruleset name, in the following format: RuleAppName/RuleAppVersion/RulesetName/RulesetVersion
Executions
The number of executions of this ruleset.
Average execution duration
The average duration of ruleset execution, which is computed by dividing the sum of all execution durations by the number of executions.
Note: Execution durations do not include the duration of the event emission process.

Errors

Status selector
This widget lets you select one or more status on which to filter the decisions. The statuses are described in Service health.

Click Apply changes for the filter to take effect and display the corresponding decisions in the dashboard. To disable the filter, click Clear form or use the Kibana user interface.

Rulesets in error
This widget displays a table that shows all the rulesets that are in error state.
Rulesets
The ruleset name in the following format: RuleAppName/RuleAppVersion/RulesetName/RulesetVersion
Executions
The number of executions of this ruleset that have failed or for which the trace could not be generated.
Errors cloud
This widget displays a cloud of the error messages that are returned during the period that is marked by the time stamp within the Kibana selected time range. The size of a message in the cloud depends on its number of occurrences. Clicking a message name filters the entire dashboard to show only the rulesets where this message occurred.

Search

At the bottom of the Kibana window, the Search widget displays all the decisions in the current time range as a table with the time stamp, the ruleset (in the same format as in the other widgets), the ruleset execution duration, the type of the trace. To filter directly on this ruleset execution, hover on its name and use the Filter this value feature provided by Kibana. Each event can be expanded to get all its contents in tabular or in JSON format.