Data Quality Thresholds
Overview
Data quality thresholds let you define acceptable levels for your data quality results. A threshold indicates when data quality is acceptable and when it needs attention, and triggers alerts when results fall below the defined value.
How thresholds relate to rules
Thresholds work alongside DQ evaluation rules. When you apply rules to your data — whether through terms or directly to attributes — DQ evaluation runs and produces results as percentages. You can set thresholds at the rule level to evaluate individual rule results, or at the monitor level to evaluate overall data quality. When results fall below the threshold, they are flagged as Critical and an alert is triggered.
How thresholds work
Thresholds work alongside DQ evaluation rules. When you apply rules to your data - whether through terms or directly to attributes - DQ evaluation runs and produces results as percentages. Thresholds give these results meaning by defining what percentage is acceptable for your use case, and trigger alerts when results fall below that level.
A threshold divides data quality results into two severity levels:
-
Critical: Results below the threshold value. These indicate data quality that needs attention.
-
Valid: Results at or above the threshold value. These indicate acceptable data quality.
For example, if you set a threshold of 80%:
-
Results below 80% are flagged as Critical.
-
Results from 80% to 100% are considered Valid.
When DQ evaluation runs and the result falls below the threshold, an alert is triggered.
Where to set thresholds
You can configure thresholds at three levels:
- DQ rules
-
Set a default threshold for a rule. This default is applied whenever the rule is used, but can be overridden for specific rule instances.
- Rule instances
-
Set a threshold for an individual rule applied to an attribute. If a default threshold is defined on the rule, it appears automatically and can be adjusted for the specific instance.
- DQ monitors
-
Set a threshold for the overall data quality result of a monitor.
Set a default threshold on a DQ rule
-
Navigate to Data quality > DQ rules.
-
Select a rule from the listing.
-
Switch to the Implementation tab.
-
In the DQ Threshold section, select Set DQ Threshold.
-
Set the percentage value.
The display shows the resulting ranges: results below your value are Critical, results at or above your value are Valid.
-
Save and publish your changes.
This threshold is applied as the default whenever the rule is used. Users can override it when applying the rule to specific attributes.
Set a threshold on a rule instance
-
Navigate to Catalog > Data catalog.
-
Select a catalog item from the listing.
-
Open a DQ monitor (primary or additional).
-
In the Profiling & Rules tab, select an applied rule to open the rule details sidebar.
-
Switch to the Configuration tab in the sidebar.
-
In the DQ Threshold section, select Set DQ Threshold.
-
Set the percentage value.
The display shows the resulting ranges: results below your value are Critical, results at or above your value are Valid.
-
Save and publish your changes.
| If a default threshold is defined on the rule, it appears here automatically. You can adjust it for this specific instance without affecting the default. |
Set a threshold on a DQ monitor
-
Navigate to Catalog > Data catalog.
-
Select a catalog item from the listing.
-
In the Overview tab, open a DQ monitor:
-
For the primary monitor, select Open DQ monitor.
-
For additional monitors, select the monitor name from the list.
-
-
Switch to the Settings tab.
-
In Processing settings, select Set DQ Threshold.
-
Set the percentage value.
The display shows the resulting ranges: results below your value are Critical, results at or above your value are Valid.
-
Save and publish your changes.
View threshold results
After DQ evaluation runs, threshold results appear across the platform:
-
DQ Results in listings: Visual indicators show threshold status at the asset level.
-
DQ View results: Detailed breakdowns include threshold violations.
-
DQ Dashboard: Summaries and trends reflect threshold status.
-
Rule instance sidebar: The Data Quality tab shows results against configured thresholds.
Was this page useful?