Microsoft Fabric monitoring in September 2026: what to switch on now

Workspace monitoring is in preview. It collects data from your Fabric items so you can analyse and monitor them. Think of it as Fabric emitting its own operational signals that you can query rather than guess at.

Because it is in preview, treat it as something to trial in a non-production workspace, not to depend on for alerting yet. It gives you a data source for analysis. It does not, on its own, give you a finished dashboard, alert rules, or a view that spans every workspace. You build that layer, or you buy it. Start now so you understand the shape of the data before it reaches general availability.

Centralized data governance in the OneLake catalog is also in preview. It aims to bring governance for your data into one place, rather than scattered across items and workspaces.

For a data quality programme, a single catalog is the foundation. You cannot enforce a rule on data you cannot find or describe. Trial the catalog, map how it fits your current governance, and keep your existing controls running while it matures. Preview means the design can still shift, so do not retire anything that already works.

This is the one to schedule. The ODBC to ADBC driver transition is Generally Available. The ADBC drivers are available now for validation. The important line is easy to miss: embedded ODBC drivers will be disabled in a future version. Any connection that relies on an embedded ODBC driver has a deprecation clock running.

1. Inventory every connection and tool that uses an embedded ODBC driver.
2. Test each one against the ADBC drivers now, while both paths still work.
3. Fix or replace anything that breaks, and record an owner for each connection.
4. Set a target date to switch before the ODBC drivers are disabled.

You do not know the exact version that disables the embedded drivers, so do not wait for it. Validate now while you still have both options. A pipeline that fails on a driver you were warned about is an avoidable outage.

Each of these items helps. None of them, on its own, gives you one view of pipeline health, refresh failures, and data quality across every workspace. Workspace monitoring is a feed. The networking API is a posture check. The catalog is governance. A platform team still has to stitch them together, then keep the stitching working as Fabric changes.

Here is a short plan for September. Four moves, in order of how quickly they pay off.

1. Use the Networking Communication Policies Admin API to audit workspace inbound and outbound policy across the tenant.
2. Trial workspace monitoring in a non-production workspace, and note what it does and does not give you.
3. Trial the OneLake catalog for governance, and map it to your data quality rules.
4. Inventory and test your ODBC connections against the ADBC drivers, and set a switch date.

If you would rather see all of this in one view than build and maintain it yourself, book a walkthrough of the Fabric Monitoring Tool with your own workspaces at /subs-services/fabric-monitoring-tool.

1. What's new in Microsoft Fabric (accessed 2026-09-18)