Reading the Logging Report¶
Every page of the Logging Report is built from the same ingredients: a tab strip, a header with freshness information, a filter bar and a set of status icons. Once you know these, every page reads the same way.
The pages¶
The tabs at the top of the report take you to the six pages.
| Page | Use it to |
|---|---|
| Overview | See the current state of every entity in every layer, in one screen |
| Entity Load History | Compare an entity's result across days, and spot entities that stopped loading |
| Landingzone, Bronze and Silver | Get run-level detail for one layer: durations, row counts and links to Fabric |
| Failed Entities | Work through everything that failed in the selected period |
A house icon in the top-left corner of every page returns you to the Overview, and the arrow next to it takes you back to the page you came from.
The header¶
| Element | Meaning |
|---|---|
| Last event time | The start time of the most recent activity in the report. Use it to judge how fresh the data is. |
| Selected period | The period the page is currently showing, matching the Period filter. |
| Environment switch | Switches between your environments, for example DVLM (development) and PROD (production). |
Check the environment first
The environment switch decides whether you are looking at your development or your production platform. If a load you expect is missing, confirm you are on the right environment before you start troubleshooting.
The filter bar¶
Below the header, each page has a row of filters. Not every page carries every filter, but the ones it does carry always mean the same thing.
| Filter | Filters on |
|---|---|
| Load Groups | The Load Planner Group that triggered the load |
| Domain Name | The domain the workspace belongs to. Labelled Domain on the Entity Load History page |
| Datasource Name | The data source the entity comes from. Labelled DataSource on the Entity Load History page |
| Source Schema | The schema, folder or container the entity comes from |
| Source Name | The table, file or object name of the entity |
| Entities With Status | Show entities where Started, Succeeded or Failed occurred at least once in the period |
| Period | The time window: 24H, 48H, 7D, 30D or YTD |
| Process Start Time | One specific process run, identified by the moment the process started |
Three things are worth knowing about the filters:
- The period is a rolling window, not a calendar period. 24H means the last twenty-four hours counted back from now, not "since midnight". YTD means since 1 January of the current year.
- The filters only list values that actually occur. If a data source does not appear in the DataSource filter, no activity was recorded for it in the retained history.
- Filters apply to the page you are on. Each tab has its own filter bar, so a selection on the Overview page does not follow you to the Bronze page.
Reset to 24H Default¶
The Landingzone, Bronze and Silver pages have a Reset to 24H Default button at the end of the filter bar. It clears the filter bar and puts the page back to its starting state: Period on 24H, every other filter on All.
Reset always returns you to Production
The button restores the environment as well, and it always restores it to PROD. If you were investigating a development environment when you pressed it, you are now looking at production data with no other visible change on the page.
Check the environment switch in the header after every reset, and set it back to the environment you were working in if you need to continue there.
The status icons¶
Almost every result in the report is expressed with one of three icons.
| Icon | Status | Meaning |
|---|---|---|
| ✅ | Succeeded | The activity finished successfully |
| ⏳ | Started | The activity reported a start and has not reported an end. Usually it is still running — but see below |
| ❌ | Failed | The activity reported a failure |
| (blank) | No activity | Nothing was recorded for this entity in the selected period |
When one icon summarises several activities — a whole layer, an entity across three layers, or a complete run — the report shows the most severe status it finds:
Failed beats Started, and Started beats Succeeded.
So a single ❌ anywhere underneath a collapsed row turns that row red. Expand the row to find which activity caused it.
What Started really tells you¶
A load reports its own progress: first that it has started, and later that it has succeeded or failed. Until that second result comes in, the report can only show Started.
Most of the time that is exactly right and the load really is in progress. But a load that never reports a result stays on ⏳ for good. That happens when:
- The run broke off before it could report. A Spark session that is lost, a timeout, or a capacity problem can end an activity so abruptly that it has no chance to send a failure. The load did fail — the report simply never heard about it.
- Someone cancelled the run. A run cancelled from the portal stops without a final status, so its entities keep showing ⏳ instead of turning ❌.
A ⏳ that is not moving is a warning sign, not progress
Judge a ⏳ against how long that load normally takes. One that is still amber long after the load should have finished is very unlikely to be running: treat it as a failure that never got the chance to report itself.
Open the run through its Monitoring URL to see what Fabric says. Fabric knows whether the run is genuinely active, cancelled or finished, where the report can only show the last result it was given.
Cancelled runs stay on Started
Cancelling a load in the portal leaves its activities without a succeeded or failed message, so they remain ⏳ for the rest of the retention window. They are not stuck and they are not still running — they simply never reported a result. Keep it in mind when a cancelled load leaves a trail of amber behind it.
Getting to the underlying detail¶
The Logging Report tells you what happened. For the why — the actual error text, the Spark log, the activity input and output — the report links straight into Microsoft Fabric.
- Detail tables contain a Monitoring URL column. Selecting the link opens the exact pipeline run or notebook run in the Fabric monitoring hub, where the error message and full run detail live.
- Right-clicking a row and choosing Drill through opens a layer page filtered to that entity. Pick the layer of the row you clicked — the menu offers all three and preselects none. See From a red cell to the run details.
- Hovering over a status icon on the Overview page shows a tooltip with the latest run details for that entity.