Version 26.10.0

Date: 10th October 2026

Data Reliability

• Separate Statuses for Execution and Data Quality Failures: The Overall Status on the Manage Policies page now shows Failed when a policy execution does not complete, and Errored when the execution completes but the quality score does not meet the policy threshold. Errored policies also display their quality score, so you can identify data quality failures without checking the Jobs page.
• Scoring Message for Combine Good & Bad Records: ADOC now displays a message when you enable Combine Good & Bad Records? on a data quality rule, stating that the rule is excluded from policy scoring and its weight is set to zero. For more information, see Data Quality Policy.
• Search, sort, and filter rules in Data Quality policies: You can now search, sort, and filter the list of configured rules in a Data Quality policy, which makes rules easier to find in large policies. You can’t sort by Threshold Config, Tags, Criteria, or Weight. For more information, see Data Quality Policy.
• Auto Anomaly Jobs on the Jobs Page: The Jobs page now includes an Auto Anomaly tab, where you can monitor Auto Anomaly policy jobs, including executions that ran before this release, and cancel jobs that are still running. For more information, see Monitoring and Managing Jobs.
• Tooltips for 30-Day Metrics (ACR-3035):Tooltips for 30-day metric columns: The 30-Day Update Count and 30-Day Profile Anomalies columns on the Discover Assets page now show a tooltip when you hover the info icon in the column header. The tooltips explain that the update count is the number of schema updates and the anomaly count is the number of anomalies found across profiling runs, both over the last 30 days. For more information, seeDiscover Assets.

Data Sources and Connectors

• Apache Kyuubi Connector Data Source: ADOC now supports Apache Kyuubi as a data source, so you can crawl Kyuubi databases and monitor them with data reliability policies using the Pushdown data plane engine. You can optionally authenticate using Keycloak (OAuth2) and retrieve credentials from a secret manager. For more information, see Apache Kyuubi.
• Secret Manager for Databricks Service Principal: Resolved an issue where Databricks connections using Service Principal authentication failed validation when the client secret came from a secret manager. For more information, see Databricks Reliability.

Dashboards

• New Dashboard Templates: ADOC now provides six new templates for creating dashboards on the Data Reliability tab: Schema Drift Dashboard, Profiling Dashboard, Monthly Data Quality KPI Report, Freshness Dashboard, DQ Freshness Check Dashboard, and Data Product Quality Dashboard. Each template in the Choose Your Dashboard Template dialog now includes a description of the data it displays. For more information, see Data Reliability Tab.
• Data Product Quality Dashboard: ADOC now provides a Data Product Quality Dashboard template that summarizes data quality across your published data products, including reliability scores, assets grouped by reliability, policy execution results, and governance coverage by policy type. You can open a details panel from a widget to see the items behind a value, and select a chart segment to filter the dashboard by that value. The Data Products widgets used in this template are also available in the widget library, so you can add them to any dashboard. For more information, see Data Reliability Tab.
• Additional Date Range Options for Dashboards: ADOC now provides more date range presets in the time filter on Data Reliability dashboards, including Last 15 minutes, Last 30 minutes, Last 1 hour, Last 3 hours, Last 6 hours, Last 12 hours, Last 24 hours, Today, Yesterday, Last 6 months, and All time. You can also select a custom start and end date and time. For more information, see Data Reliability Tab.

Pipelines

• Resizable Details Panel for Pipeline Runs: You can now resize the details panel below the lineage graph on the Pipeline Run Details page by dragging its top edge, or collapse it to give the lineage graph the full view. A taller panel gives the Timeline more space to show task durations and parallel execution in large pipelines. For more information, see Pipeline Run Details.
• Horizontal and Vertical Pipeline Graph Layouts: You can now switch the lineage graph on the Pipeline Run Details page between a vertical (top-to-bottom) and a horizontal (left-to-right) layout. The horizontal layout shows more nodes on wide screens and reduces scrolling in large pipelines. For more information, see Pipeline Run Details.
• Persistent Dependency Highlighting in Pipeline Graphs: You can now Command+click (macOS) or Ctrl+click (Windows) nodes in the lineage graph on the Pipeline Run Details page to keep them and their directly connected nodes highlighted while you pan and zoom. To focus on the selected nodes alone, click Isolatethe selected nodes or hold S.For more information, see Pipeline Run Details.
• Task Retry Counts in Pipeline Graphs: Task nodes in the lineage graph now show how many times a task was retried before it succeeded, for Airflow and AWS Step Functions pipelines. Select a task's status icon to view its events. For more information, see Pipeline Run Details.
• Run-Level Links for Dataset-Triggered Airflow DAGs: ADOC now links the Airflow run that updated a dataset to the dataset-triggered DAG run it started, so you can open the exact upstream or downstream run from the lineage graph instead of the latest run. ADOC confirms these links on Airflow 2.11 and later, and infers them from the dataset and execution interval on Airflow 2.10. Links are created only for runs ingested after the upgrade. For more information, see Link Dataset-Triggered Airflow DAG Runs.

Security & Authorization

• Automatic Group Assignment from SAML: ADOC adds users to matching ADOC groups based on the group membership in the SAML assertion, without SCIM. Changes take effect at the user’s next sign-in. For more information, see SSO and User Groups.

• Resolved an issue where selecting an aggregate rule, such as a Uniqueness or Metric Check rule, in the Execution Summary displayed the message "Cannot filter bad rows in Uniqueness rule," which did not explain why sample rows were unavailable. ADOC now displays a message explaining that sample row-level results are not available for aggregate rules because these rules evaluate a metric across the entire dataset.
• Resolved an issue where Reconciliation policies stayed in the RUNNING state and couldn't be stopped or rerun. The Spark driver moved to ABORTED, but the driver pod didn't terminate and the job status wasn't updated, and the Stop action failed with a 500 Internal Server Error.
• Resolved an issue where Incremental Profiling jobs on the Pushdown engine failed or stayed stuck in the SUBMITTED state when the incremental strategy column wasn't also chosen under Select Columns.
• Resolved an issue where an asset's Reliability and Data Quality scores showed different values on Discover Assets, Asset Overview, the Policies tab and the homepage, with no clear scope, calculation or refresh details to explain the difference.
• Resolved an issue where ADOC didn't recognize Airflow Dataset scheduling when a DAG used a single Dataset object (schedule=dataset) instead of a list (schedule=[dataset]), so the upstream Dataset lineage was missing for the consuming DAG.
• Resolved an issue where generating sample bad data records for a Data Quality policy failed with a 500 Internal Server Error on the GET /api/rules/sample/result/execution-result-<ID> endpoint.
• Resolved an issue where crawling JSON files on an Amazon S3 data source failed with a URISyntaxException when the file path contained a wildcard (*).
• Resolved an issue where Reconciliation rule descriptions didn't display correctly.
• Resolved an issue where the Data Quality execution results API didn't always return the actual rowsFailed count for rule items. The count is now returned whether the rule passes or fails.
• Resolved an issue where the Rules listing page didn't show the owning policy by default, so rules from different policies that shared the same name, asset, column, rule type and dimension couldn't be told apart.
• Resolved an issue where switching a Data Quality rule's threshold type from Absolute to Relative and back to Absolute reset the success and warning thresholds to default values without any warning, discarding the values the user had set.
• Resolved an issue where Data Quality policies failed with a PatternSyntaxException (Invalid regular expression) during execution.
• Resolved an issue where the YuniKorn scheduler pod had no liveness or health probe. The scheduler could silently stop processing new pods while still reporting as healthy, leaving policy and profiling jobs stuck in Pending or Submitted until the pod was restarted manually.
• Resolved an issue where Domain and Tag filters behaved inconsistently on dashboards.
• Resolved an issue where Pipeline Automation showed an incorrect label and default value when a Data Quality or Reconciliation policy was selected.
• Resolved an issue where Profiling, Sample Data and policy runs on the Spark engine failed with an UNRESOLVED_COLUMN.WITH_SUGGESTION error for Databricks tables whose column names contain a period (for example, Ref. Document No).
• Resolved an issue where the DQ Score by Asset Name and Policies Configured dashboard widgets showed errors and the Top Performing Tables By Quality widget listed the same asset more than once.
• Resolved an issue where expanding a related downstream pipeline node from a past pipeline run showed the downstream pipeline's latest run instead of the run triggered by the selected upstream run.
• Resolved a performance issue where the Policy Executions dashboard widget loaded slowly.