Overview

Each monitor run ends with a display of its results, which you can see in the Runs tab of the Monitor Details Page.

Analyze Results

If a monitor fails (detects anomalies), you can preview the rows that didn't meet the monitor criteria using Failing Rows. For machine learning (ML)-based monitors, you can also display results on graphs.

Failing Rows

For all monitors that detect anomalies on specific rows (for example, Format Validation), you can preview Failing Rows. To display the failing rows, click the Show failing rows button in the Runs tab.

Graphs

For ML-based monitors, you can analyze results with graphs. Those are displayed both in the Overview section of a monitor for the last run and in the Runs section for a deep dive on a specific past execution. To get those graphs, click the Show details icon on any line.

Analysis

• Green area: The confidence band computed by the ML model represents the expected behavior of the data.
• Green dotted curve: Actual behavior of the data.
• Red dots: Anomalies, when the actual-data curve is outside of the confidence band.