• Announcement
• New and improved
• Fixes

Release Information

• Release date of Data Quality & Observability Classic 2025.04: April 28, 2025
• Release notes publication date: April 2, 2025

Announcement

New and improved

Platform

• The Platform Path of the SQL Assistant for Data Quality feature now uses the Gemini AI model version gemini-1.5-pro-002.

Important To continue using the Platform Path of the SQL Assistant for Data Quality feature, you must upgrade to Data Quality & Observability Classic 2025.04.

• FIPS-enabled standalone installations of Data Quality & Observability Classic now support SAML authentication.
• The DQ agent now automatically restores itself when the Metastore reconnects after a temporary disconnection due to maintenance or a glitch.

Jobs

• We are pleased to announce that Oracle Pushdown is now available for preview testing.
• When you add a timeslice to a timeUUID data type column in a Cassandra dataset, an unsupported data type error message now appears.
• Dremio, Snowflake, and Trino Pullup jobs now support common table expressions (CTE) for parallel JDBC processing.
• You can now archive the break records of shapes from Trino Pushdown jobs.
• Link IDs for exact match duplicates are no longer displayed on the findings page.
• We improved the insertion procedure of validate source findings into the Metastore.

Rules

• When you add or edit a rule with an active status, SQL syntax validation now runs automatically when you save it. If the rule passes validation, it saves as expected. If validation fails, a dialog box appears, asking whether you want to continue saving with errors. Rules with an inactive status save without validation checks.
• When a rule condition exceeds 2 lines, only the first 2 lines are shown in the Condition column of the Rules tab on the Findings page. You can click "more..." or hover over the cell to show the full condition.

Profile

• The Profile page now correctly shows the top 5 TopN and BottomN shape results for Pullup jobs.
• When there are only two unique string values, the histogram on the Profile page now shows them correctly.

Findings

• To improve page alignment across the application, the Findings page now has a page title.

Alerts

• Rule score-based alerts now respect rule tolerance settings. Alerts are suppressed when the rule break percentage falls within the tolerance threshold.

Dataset Manager

• If you don't have the required permissions to perform certain tasks in the Dataset Manager, such as updating a business unit, an error message now appears when you attempt the action.

Scorecards

• You can now delete scorecards with names that contain trailing empty spaces, such as "scorecard " and "test scorecard ".

Integration

• You can now search with trailing spaces when looking up a community for tenant mapping in the Data Quality & Observability Classic and Collibra Platform integration configuration.

APIs

• You now need the ROLE_ADMIN or ROLE_ADMIN_VIEWER role to access the /v2/getdatasetusagecount API endpoint.

Fixes

Platform

• We have improved the security of our application.
• The data retention process now works as expected for all tenants in multi-tenant instances.
• Permissions errors for SAML users without the ROLE_PUBLIC role are now resolved. You no longer need to assign ROLE_PUBLIC to users who already have other valid roles.

Jobs

• Native rules, secondary dataset rules, and scheduled jobs on Databricks Pullup connections authenticated via EntraID Service Principal now run as expected.
• Pushdown jobs scheduled to run concurrently now process the correlation activity correctly.
• Pushdown jobs with queries that contain new line characters now process correctly, and the primary table from the query is shown in the Dataset Manager.

Findings

• The confidence calculations for numerical outliers in Pullup jobs have been updated for negative values. Positive value confidence calculations and Pushdown calculations remain unchanged.

Alerts

• Conditional alerts now work as expected when based on rules with names that start with numbers.

Integration

• The default setting for the integration schema, table, and column recalculation service (DQ_DGC_MAPPING_STATS_SCHEDULER_ENABLED) is now false, reducing unnecessary database activity. You can enable the service or call it through the API when needed.

Important From versions 2024.11 to 2025.03 of Data Quality & Observability Classic, if you don't want queries from the recalculation of mapped and unmapped stats of total entities to run, set the DQ_DGC_MAPPING_STATS_SCHEDULER_ENABLED to false in the owl-env.sh or Web ConfigMap. You can keep DQ_INTEGRATION_SCHEDULER_ENABLED set to true.
For more information, go to the Collibra Support center.

• The Quality tab for database assets is now supported in out-of-the-box aggregation path configurations.
• The auto map feature now correctly maps schemas that contain only views.
• Dimension configuration for integration mapping no longer shows duplicate Data Quality & Observability Classic dimensions from the dq_dimension table.

API

• When you copy a rule with an incorrect or non-existent dataset name using the v3/rules/copy API, an error message now specifies the invalid dataset or rule reference. This prevents invalid references in the new dataset.