Adding Custom Data to Insights

Important:

• For deployments exceeding 75 processes and/or queues with custom data added per tenant, you must either add your desired process and queue tables to an allowlist or Enable Table Grouping.
• A configurable alert is displayed warning you when you are approaching this limit.
• Table grouping may also need to be enabled based upon the total number of such tables across all tenants. Review the Hardware and Software Requirements for additional details.
• The number of columns in process and queue tables is limited to 40.

Workflow Design for Insights

Important: Unique case-insensitive names must be used for all custom variables (i.e. do not use successfulTransaction and SuccessfulTransaction). If your variables are not unique, you must set Insights to be case insensitive.

When designing your automation workflows intended for consumption and analysis in Insights, there are changes to consider in the best practice approach used to date.

You do not need to "tag" individual items and/or data points to ensure their availability for analysis in Insights. For all processes and/or queues with custom fields, a custom table is automatically generated with those fields available.

The following activities can be used in designing your automation workflows to capture the desired data:

• Add Log Fields - all added log fields are ingested as a custom table in Insights. See [here] (doc:adding-custom-data-to-insights#logs) for more details.
• Add Queue Item - all data added via the ItemInformation property are ingested as a custom table in Insights. See here for more details.
• Set Transaction Status - all data added via the Analytics and Output properties are ingested as a custom table in Insights. See here for more details.

For complete details about all data ingested by Insights, see the Insights Data Model page.

Note:

For those workflows that you want to analyze in Insights it is recommended to log all desired data at the end of each workflow.

Where previously you would use the RemoveLogFields activity at the end of any individual sequence to remove logs not needed in the rest of the workflow, removing all custom logs as such results in the process and data not being ingested into the Insights database.

Logs

There are 2 methods to leverage logs for sending data to Insights:

• Log message - enables you to add a custom string message that is sent to Insights, in the form of a log event, in case of an error in your workflow.
• Add Log Fields - enables you to add custom variables to the default Insights dashboards.

When Insights detects that a log message has custom log fields added via the Add Log Fields activity, a new table is created in Insights containing those custom fields with the title "Process-<your process name>”. These custom fields are extracted from the raw message via a custom JSON parser and added to the table in the RawMessage_CustomFieldName format.

The following data types for custom fields are supported: Integer, String, Boolean, Double, and Float.

Note: The custom tables are not joined to the OOTB data schema, so cross-table widgets will not be possible unless the relationship is created. For such customization, please reach out to your UiPath representative.

Insights receives the last log event only for successful processes. If you use the Remove Log Fields activity, Insights will not be able to capture the custom fields as they will be removed by the workflow prior to the process completing.

This also means that transactional processes that leverage logs for capturing data per transaction will not send the data for each transaction log to Insights. If you are using a process that iterates through transactions and wish to store custom data for each, you should follow UiPath best practices and use queues. Queues support sending custom variables for every queue item to Insights.

Queues

Every queue item has three sections in which you can store custom variables. Upon creating a queue item via the Add Queue Item activity, a field called ItemInformation becomes available. There, you can store any desired custom variables. The supported data types for all queue data are the same as for custom logs: Integer, String, Boolean, Double, and Float.

To add custom variables to a queue item once it has finished processing, leverage the Set Transaction Status activity. It contains two Inputs in which you can store variables: Analytics, and Output.

If custom fields are detected by Insights in the queue, a custom table is created in the data model called Queue-"your queue name”. The fields will be identified via the input they were stored in:

• Variables stored in Analytics will be named “AnalyticsData_<yourfieldname>”.
• Variables stored in Output will be named “Output_<yourfieldname>”
• Variables stored in ItemInformation will be named “SpecificData_<yourfieldname>”.

Note: The queue custom tables are not joined to the OOTB data schema, so cross-table widgets will not be possible unless the relationship is created. For such customizations, please reach out to your UiPath representative.