The short version: the Monitor hub is where you start when someone asks “did the job run?” in Microsoft Fabric. One pane, all item types, no clicking through every workspace.
One caveat up front. The Monitor hub is currently in preview. This post is based on the Microsoft documentation (read on 9 October 2026), not on a dedicated test workspace. Where I give an opinion, I say so.
What it is
The Monitor hub brings job execution health, alerts, capacity, agents and application telemetry together. Before, you opened the pipeline, then the notebook, then the dataflow. Now you start in one list.
Microsoft organizes it in five pages:
- Job runs: runs across the items in your workspaces.
- Applications: application health and usage.
- Capacity and cost: capacity health and consumption.
- Agents: agent activity and health.
- Alerts: who gets notified about job failures and other job events.
What does not change: you only see items you have permission to view. Any Fabric user can open the hub, but it is not an admin back door.
Job runs: the page you will use most
Open Monitor in the navigation pane, switch to the new experience if you still see the classic view, and select Job runs.
Each row is an item with its last run status (In progress, Succeeded, Failed, Canceled) and a short history of the last five runs. That alone answers most daily questions.
Practical habits that make it useful:
- Filter on Status = Failed to create a worklist.
- Filter by Item type or Workspace when you own a part of the estate.
- Sort on Average duration to find the slow ones.
- Save the result with Manage view > Save view. Your failed-pipelines view is then one click away.
Two details that catch people out. The page loads data once and does not refresh continuously, so use the reload icon. And keyword search only checks the items loaded on the current page, so use the drop-down filters first.
Select a run and a side pane opens with run history and advanced details. For pipelines and dataflows the full detail page adds the execution flow, an activity table and a Gantt chart. You can jump to the item’s authoring page from the same list.
Alerts: free email or Activator
This is the distinction I would not skip. There are two alert types, managed together but separate:
| Failed scheduled job emails | Activator-based rules | |
|---|---|---|
| Good for | An email when a scheduled job fails | More events (Started, Succeeded), richer notifications, automated actions |
| Cost | Free | Activator capacity usage |
| Needs | Contributor on the workspace or Write on the item | Owner or Contributor, because Workspace Monitoring and Activator items are created |
Recipients of the free emails must be users or groups in your Microsoft Entra tenant (guests included). Direct external email addresses are not supported.
In practice: start with the free failure emails on every scheduled item that matters. Add Activator rules only where “it failed” is not enough, for example when something should happen automatically.
Failed pipeline? Let the agent look first
For failed pipeline runs, the hub offers an entry point to the read-only Operations agent (preview) in Real-Time Intelligence. It reviews the run and suggests likely root causes without changing the pipeline.
This sounds small, but it shortens the step between “red” and “why”. I have not tried it yet, so I will not claim how good the answers are. Setup, prerequisites and limitations are in Microsoft’s operations agent documentation.
My take
The interesting part is not the list. It is that failures, alerts and capacity now sit in one place, so the question “who looks at this?” becomes answerable.
I am not fully convinced yet that preview-level monitoring should be your only safety net. Keep your existing alerting for critical loads until this is generally available, and check the preview terms.
Try it this week
- Open Monitor > Job runs, filter on Failed, save the view.
- Configure Schedule failure emails for your five most important scheduled items.
- Decide per item whether an Activator rule is worth its capacity cost.
If you are sizing capacity as well, read Plan Microsoft Fabric capacity from workload patterns. If a table looks stale after a green run, see Fix missing or stale Lakehouse tables in the Fabric SQL endpoint.
Which one do you want first: better failure visibility, or earlier warnings?
Sources
- What is Monitor hub in Microsoft Fabric?, accessed 9 October 2026
- Monitor job runs in the Monitor hub, accessed 9 October 2026
- Set up job alerts in the Monitor hub, accessed 9 October 2026
Loading comments…