Enable workspace monitoring in Microsoft Fabric

Workspace monitoring gives you log-level visibility for all items in a workspace, including pipelines. It stores execution logs in a monitoring eventhouse so you can query and analyze them using KQL (Kusto Query Language).

Enable workspace monitoring (Preview)

1. In your Fabric workspace, select Workspace Settings, then select the Monitoring tab.

   

2. Turn on Log workspace activity. This creates an eventhouse in your workspace to store logs. Fabric also creates a read-only KQL database inside the eventhouse for monitoring data.

   

3. Go to the KQL database. You can find it in the Monitoring database link within the monitoring settings or in your workspace.

4. In the KQL database, the ItemJobEventLogs table captures pipeline-level events from your workspace. Logs include pipeline name, run status, timestamps, and system diagnostics.

   

Troubleshooting missing monitoring tables

When new workspace monitoring tables become available and you don't see them in your existing monitoring eventhouse, recreate workspace monitoring for the workspace by deleting the previous eventhouse and re-adding it to your workspace. Recreating the monitoring eventhouse provisions the latest monitoring schema and ensures newly available tables are visible. Be sure to save any custom Kql queries before removing the existing monitoring eventhouse. The monitoring experience is backed by a monitoring eventhouse and monitoring Kql database that are created when workspace monitoring is enabled.

Review logged data

The ItemJobEventLogs table includes pipeline-level (L1 across Data Factory items) execution data:

• Pipeline name
• Run status (Success/Failed)
• Start and end timestamps
• System diagnostics

For pipeline activity-level (L2) monitoring, you can find more logs in the FabricDataPipelinesActivityRunsLogs table:

• Pipeline name and run identifier
• Activity name and type
• Activity status (In Progress, Succeeded, Failed)
• Activity start and end timestamps
• Execution duration
• Failure type

Query logs

1. Go to the monitoring eventhouse.
2. Use KQL queries to analyze:
   • Success and failure trends
   • Performance metrics

Here's an example query:

ItemJobEventLogs 
| where ItemKind == "Pipeline" 
| summarize count() by JobStatus 

Use query logs to create an alert for workspace-wide pipeline failures

Use a KQL Queryset to detect pipeline failures across the workspace. Here is an example query that returns recent failures:

ItemJobEventLogs 
| extend SecondsAgo = datetime_diff('second', now(), ingestion_time()) 
| where JobType == 'Pipeline' and JobStatus == 'Failed' 
| where SecondsAgo <= 540 
| order by Timestamp desc 
| project Timestamp, ItemName, WorkspaceName, JobStartTime, JobEndTime, JobStatus 

ItemJobEventLogs schema

The following table describes the schema of ItemJobEventLogs (item-level, L1):

FabricDataPipelinesActivityRunsLogs schema

The following table describes the schema of FabricDataPipelinesActivityRunsLogs (activity-level, L2):

Best practices

• Use workspace monitoring for deep analysis and custom reporting.
• Combine workspace monitoring with the monitoring hub for quick operational checks across workspaces.

Known limitations

• Error details and diagnostics aren't supported.

Related content

• How to monitor pipeline runs
• How to monitor pipeline runs in Monitoring hub