Create a Discrete Analysis Event
|
Before you create the event, ensure that the model on which you are creating the event is enabled.
|
Complete the following steps to create a discrete analysis event:
1. From the left navigation pane, select Analysis Events, and click New.
The Create New Event window is displayed.
2. In the Source Type (Required) list, select one of the following options:
◦ Thing
◦ Thing Template
◦ Thing Shape
3. In the
Source (Required) list, click
to search for and select the actual instance of the event source.
4. In the Event (Required) list, select the event on which you want to trigger the analysis.
5. In the Property (Required) list, select the name of the event source property, if the event is property-specific.
6. In the Alert Name list, select the name of the alert, if the event is alert-specific. This field is available only if the source thing contains an alert property.
7. In the Provider Name (Required) list, select the analysis provider.
8. In the Model Name (Required) list, select the analysis model on which you want to execute the analysis.
| Only those models that are enabled are available in the list. |
9. Clear the Run as Simulation check box, if selected.
10. In the Retain Job History list, select one of the following options to set how you want to store jobs triggered for this analysis event on the Analysis Jobs page:
◦ None — Saves none of the analysis jobs.
◦ Failed — Saves only failed analysis jobs.
◦ All — Saves all analysis jobs.
By default, None is selected.
| If while creating your analysis event, you configure your event not to save jobs, and then later you want to save all the jobs or only failed jobs, you can update the analysis event. To do this, see the “View an Analysis Event” section in Manage an Analysis Event. |
11. In the Collect trailing data for list, select one of the following options to set how you want to collect trailing data:
| The Collect trailing data for option is available only if the analysis model has the Inputs to Results Relationship selected as Undefined, Many to Many, or Many to One. For more information about the different options available, see Create an Analysis Model. |
◦ None — Select this if you do not want to collect trailing data for the analysis event.
◦ Time Window — Select this, and then specify the time in days, hours, minutes, seconds, or milliseconds for which you want to collect the trailing data.
◦ Lookback Count — Select this, and then specify the last ‘X’ number of instances that you want to collect trailing data for.
12. If you want, select the Data Sampling Frequency check box, and enter the frequency at which you want to collect the trailing data.
The following table provides examples with possible combinations of the Collect trailing data for and Data Sampling Frequency options.
Collect Trailing Data for | Data Sampling Frequency | Example |
---|
Time Window | No | If you have specified the value of Time Window as 10 minutes, and data for 100 minutes is available, values of only the last 10 minutes are collected. |
Yes | If you have specified the value of Time Window as 10 minutes, and the Data Sampling Frequency is set to 5 minutes, then data for 10 minutes is collected. This corresponds to 1 record every 5 minutes. |
Lookback Count | No | If you have specified the value of Lookback Count as 10, and 100 values are available, values of only the last 10 instances are collected. |
Yes | If the Lookback Count is set to 10 and the Data Sampling Frequency is set to 5 minutes, then data of the last 50 minutes is collected. This corresponds to 1 record every 5 minutes. |
13. Click Save to save the analysis event.
After you create an analysis event, you must map input data and result data to properties. For more information, see
Map Event Data to Properties.