Monitor Data Loads using C3 AI Studio
C3 provides robust monitoring tools to manage data loads. This section describes C3 AI Studio's no-code tools to inspect any errors that may occur during the process.
Open data load monitoring tools
To begin monitoring your data load, first navigate to C3 AI Studio and select Applications. Choose the application you want to monitor, and then select Data Integration. Access the necessary tools by selecting the following tabs: Overview, Volume Chart, Quality Heatmap, and Status Details. Please note that it may take a few seconds for the charts to load.
Identify and inspect errors
In the Status Details tab, review the data load process for any errors. You can identify failed records by filtering for "Failed" rows using the filter panel or by navigating to the second page where failed records might be listed. To inspect specific errors, select the ID of a failed row, and in the side panel that appears, select View Errors to get more detailed information.
Analyze error messages
In the modal that appears, examine the error message. For example:
Error while transforming: {
Transformation Type: SourceWindTurbineVibration-WindTurbineVibration,
Target Type: WindTurbineVibration,
errorMsg: Transformation failed: WindTurbineVibration.measurementDate.dateTime(measurement_date)
Error Msg: Invalid value '2023-50-01 01:01:00' for datetime: at 50-01 01:01:00 }Inspect the raw source file
To inspect the raw source file associated with an error, select the Source Status button in the side panel. Then, under View Source File," select the Source file to view the raw data. Within this file, search for the problematic entry (for example, a date like 2023-50-01) to identify the invalid record.
Understand error handling
It is important to understand how error handling works in the C3 Agentic AI Platform. The platform automatically catches errors such as invalid datetime formats, performing validation on a record-by-record basis. This approach prevents large data loads from failing entirely due to a few bad records.
Determine next steps
After identifying the issue, determine the appropriate next steps. If the error is due to an invalid record (like an incorrect datetime), no further action may be necessary. However, if the error relates to a transform definition or other Type definitions, use the error message to troubleshoot and resolve the issue, then re-load the data as needed.