Fabric Workspace Monitoring: What It Actually Shows You

Ever had a Fabric pipeline hang for no obvious reason, or watched a semantic model refresh crawl with no explanation? Fabric Workspace Monitoring is built for exactly that. Flip one toggle in workspace settings and Fabric spins up a read only Eventhouse, streaming execution logs, query text, CPU consumption, and error data into KQL tables you can query directly. Nothing to install, no separate platform. Microsoft’s own steps to enable it are three clicks, Workspace Settings, the Monitoring tab, then turn it on.

What Fabric Workspace Monitoring actually logs

Coverage isn’t uniform yet, but it’s grown quickly.

• Semantic models – trace events: commands, DirectQuery calls, errors, execution metrics.
• Data pipelines – run status in ItemJobEventLogs, plus activity level detail in FabricDataPipelineActivityRunsLogs, both covered in Microsoft’s pipeline monitoring guide.
• Eventstreams – node status, throughput, and error counts, currently in preview per the Eventstream monitoring docs.
• Mirrored databases – replication status, table changes, and latency.
• Eventhouse – every KQL query and command run against it.

It all lands in one place, so you can correlate across item types, did the Copy job finish before the semantic model refresh kicked off? Microsoft’s Fabric samples repo has starting queries.

Things to look out for

Per Microsoft’s Workspace Monitoring overview, you can only run Workspace Monitoring or Log Analytics in a workspace, never both. Switching from Log Analytics means deleting that configuration first, then waiting a few hours before Workspace Monitoring will turn on. Deleting the monitoring Eventhouse itself carries a similar wait, about 15 minutes before you recreate it.

What Fabric Workspace Monitoring costs

Workspace Monitoring runs on two Fabric items, an Eventstream and an Eventhouse, both billed against your capacity. Microsoft confirmed the rates in their billing announcement.

• Eventstream – 0.22 CU/hour, plus 0.342 CU/hour per GB of traffic.
• Eventhouse – 1 CU/hour per active v-core for compute.
• Eventhouse storage – billed separately across OneLake cache and standard storage.

A single workspace won’t notice the cost. Dozens of workspaces, and it adds up, which is why one monitoring instance covering multiple workspaces sits on Microsoft’s own roadmap.

Turning Fabric Workspace Monitoring into alerts, not just logs

Logs sitting in a KQL database are only half the job. Someone still needs to get told when things break. Microsoft’s guide to alerting on pipeline runs lays out the pattern.

1. Confirm workspace monitoring is enabled, so ItemJobEventLogs is populated.
2. Create a KQL Queryset against the monitoring Eventhouse, filtering for what you care about, for example recent pipeline failures.
3. Run the query, then select Set alert from the ribbon.
4. Point the alert at an Activator, with an email or Teams action. Or trigger a Pipeline among other options!
5. Save and enable it.

A simple query filters ItemJobEventLogs down to JobStatus == "Failed" within a rolling window. Saved as an alert, it covers every pipeline in the workspace, not just one.

Two things to check first, both from Microsoft’s Activator queryset documentation. Activator needs to poll more frequently than your query’s time window, or you’ll miss events. And if Activator gets paused or disabled, alerts stop firing silently. A query with joins or heavy transformations, running every few minutes, also keeps the Eventhouse always on, feeding back into the cost above.

Should you turn on Fabric Workspace Monitoring?

If you’re troubleshooting Fabric items one click at a time, yes. Pair it with the templates from the fabric-toolbox workspace-monitoring-dashboards repo for a working monitoring UI in about ten minutes, though these are community accelerators.

It won’t catch everything. For tenant wide visibility across every workspace, that’s still the admin monitoring workspace or the Monitoring hub. But for seeing what’s happening in your workspace right now, Fabric Workspace Monitoring works well.

For more Fabric related content visit Blogs – Purple Frog Systems