Catalog collector release notes

Microsoft Fabric collector: Added support for harvesting EventStream resources, enabling lineage tracking for Fabric’s real-time streaming data pipelines.

dbt collector: Corrected resource identifier generation for tables to match BigQuery collector format, ensuring proper lineage connections between dbt models and their underlying BigQuery tables.

OpenAPI collector: Added support for OpenAPI 3.1’s updated nullable type syntax, ensuring schemas using the latest specification version are parsed correctly.

SSIS collector: Fixed parsing of SSIS dataflow tasks to handle a wider variety of schema and table name patterns in SQL queries.

Redshift collector: Added support for harvesting lineage from stored procedures, improving visibility into procedural data transformations.

ServiceNow collector: Improved pagination handling to ensure results are fully harvested when total counts do not equal 100, preventing missing resources in larger result sets.

Tableau collector: Fixed a defect affecting sheet-to-dashboard associations, ensuring relationships are harvested accurately.

Snowflake and Databricks collectors: Fixed relationship mapping so that database tables are correctly associated with their parent database, ensuring accurate hierarchy and navigation in the catalog.

Microsoft Fabric collector: Added support for harvesting Environments in Microsoft Fabric, expanding visibility into Fabric workspace configurations.

Oracle collector: Added support for harvesting lineage from CREATE TABLE statements by querying historical tables, improving lineage coverage for table creation workflows.

OpenAPI collector: Fixed an issue where restrictive server configurations caused specification requests to be rejected.

Microsoft SQL Server collector: Resolved an issue where SQL preprocessing of TOP statements did not work correctly in some cases.

Microsoft Fabric collector: Fixed issues with Lakehouse Delta tables to ensure they are cataloged under the correct parent folder and to prevent duplicate resources from being created.

SSIS collector: Added support for harvesting variables and their references, improving lineage and package-level context.

Microsoft SQL Server collector: Now harvests table descriptions from extended properties, enriching table-level documentation.

Databricks collector: Renamed the CLI option --http-path to --compute-resources-url for clearer configuration of compute connectivity.

Snowflake, Redshift, Databricks, Oracle, PostgreSQL, Teradata, MySQL, Db2, Netezza, SQL Server, SAP HANA collectors: Added support to exclude system functions across all JDBC collectors, helping reduce noise in harvested metadata.

Microsoft Fabric collector:

Sensitive data classification feature: Improved sensitive data classification for collectors. On-premise collectors now support Microsoft Presidio classification services in addition to Private AI.

Tableau collector: Fixed embedsView relationships for unpublished Tableau views and dashboards, ensuring these relationships are harvested correctly.

Azure Data Lake Storage Gen2 collector: Fixed an error that occurred when harvesting a non-hierarchical storage account, improving compatibility and stability.

Snowflake collector: Added lineage resolution support for QUALIFY clauses, LATERAL FLATTEN constructs, and subqueries in previously unsupported contexts, including WHERE, JOIN ON, and ORDER BY.

Oracle collector: The collector now harvests public synonyms, expanding visibility into shared database objects.

SSIS collector: Added support for harvesting stored procedure references, improving lineage completeness.

Tableau collector:

Snowflake collector: Fixed an issue where some SQL statements took an excessive amount of time to parse during lineage harvesting.

SQL Server collector: Prevented attempts to harvest table size metadata for views that do not support this operation.

Databricks collector: Added a safeguard to prevent SQL execution for excluded databases during lineage collection.

Oracle collector: Fixed harvesting of statistic indexes, ensuring they are correctly captured.

ServiceNow collector: Introduced a new collector that harvests metadata for scoped applications, tables, fields, views, and data fabric tables from a ServiceNow instance.

Dremio collector: The API server (--api-server) option is no longer required. When it is not provided, the collector does not harvest the API-based lineage and dataset information.

Snowflake, Redshift, Databricks, Denodo, Oracle, PostgreSQL, Teradata, MySQL, Db2, Netezza, SQL Server, SAP HANA collectors: Improved memory usage during the harvesting of column statistics, enhancing stability for large datasets.

Tableau collector:

Snowflake collector: Updated SQL parsing behavior to ensure parsing terminates correctly when a timeout occurs, and made the timeout value configurable to better handle complex or long-running queries.

Salesforce collector: Added support for grouping and aggregation columns, expanding metadata coverage for analytical queries.

Databricks collector: Added support for harvesting jobs from external workspaces, enabling visibility into cross-environment job metadata.

Azure Data Factory collector: Now harvests SQL queries used in activities, providing richer lineage and operational context.

Dremio collector:

Microsoft Fabric collector: Fixed an issue where some files were not being harvested, ensuring more complete metadata collection.

QlikSense collector: Updated logic to exclude writing the API key in the TTL file

Fabric collector: Fixed a modeling issue where Lakehouse tables were incorrectly displayed as folders, ensuring proper representation of table structures

Microsoft SQL Server collector: The collector now harvests column comments from the MS_Description extended property, improving metadata detail and context.

SQL Server Integration Services (SSIS) collector: Added support for lineage when Oracle is used as a source or target, expanding cross-platform lineage coverage.

Azure Data factory collector: Fixed an issue where lineage relationships were missing when a Snowflake credential had an assigned default role, ensuring complete lineage capture.

OpenAPI collector: Added a new configuration option --environment-qualifier. When provided, this option ensures that harvested resources are unique and distinct for that environment qualifier, improving multi-environment cataloging.

Power BI collector: The collector now harvests calculation groups and calculation items, expanding metadata coverage for Power BI semantic models.

dbt core and dbt cloud collectors: Fixed an issue where tags were represented as strings instead of relationships. Tags are now properly modeled as relationships in the catalog graph.

Power BI collector: Fixed an issue where descriptions in parameter definitions were incorrectly appended to parameter values.

Databricks collector: Added support for harvesting metrics views, expanding metadata coverage within Databricks. A new optional parameter, --enable-metric-views, is now available to enable this feature.

Power BI collector: Added support for Power BI–BigQuery column lineage.

Informatica CDI collector: Fixed an issue that caused incomplete harvesting in large CDI instances by ensuring the collector continues harvesting beyond the Informatica CDI API’s default pagination limit.

dbt Cloud and dbt core collectors: Corrected tag representation so that tags are now modeled as resources rather than strings, improving metadata consistency and usability.

OpenAPI collector: Added handling for schema references and array item references, improving completeness of OpenAPI metadata harvesting.

Power BI collector: Now harvests endorsement details for reports, datasets, and dataflows, providing better visibility into trusted and certified content.

Databricks collector: Added support for harvesting materialized views, expanding metadata coverage within Databricks.

SQL Server collector: Now harvests database synonyms, improving understanding of object relationships and dependencies.

Tableau collector: Updated logic to associate published data sources with the project that contains them, improving metadata organization and context.

SSRS collector: Now harvests the external URL for linked reports, providing direct traceability to linked content.

Snowflake, Redshift, Databricks, Denodo, Oracle, PostgreSQL, Teradata, MySQL, Db2, Netezza, SQL Server, SAP HANA collectors: Updated lineage logic to represent INSERT stored procedures as the triggering agent in lineage relationships, improving lineage accuracy.

Informatica collector: Fixed an issue that limited harvesting to a maximum of 4,200 mapping tasks. The collector now processes all available mappings.

Databricks collector: Fixed handling of non-alphanumeric characters in identifiers. Resolved issues with the creation of invalid identifiers.

dbt core and dbt cloud collectors: Fixed a case mismatch issue when coining database IRIs for Snowflake accounts, ensuring consistent IRI generation.

SSIS collector: Corrected package lookup logic to prioritize the Package GUID, using the Version GUID only when multiple versions share the same Package GUID.

Talend collector: Fixed a classification issue where Talend Jobs were incorrectly identified as activities instead of agents.

Snowflake collector: Implemented a workaround for a Snowflake JDBC driver defect that was triggered by unexpected datatypes, preventing runtime errors.

SQL Server collector: The collector now harvests table size for tables in non-dbo schemas, ensuring complete table metadata coverage.

Qlik Sense collector: Added lineage between Qlik Sense apps and datasets, providing greater visibility into data flow and dependencies.

Marquez collector: Added column-level lineage, offering more granular insight into data transformations and relationships.

Databricks collector: Fixed the creation of invalid agent IRIs, ensuring accurate and consistent identifier generation.

Qlik Sense collector: Resolved a memory leak related to closing WebSocket connections, improving stability and performance.

AWS Glue collector: Added handling for large object size properties, preventing errors during metadata harvesting for large datasets.

Marquez collector: The collector now harvests job run history. A new optional parameter, --job-run-count, is now available to enable this feature.

Tableau collector: Fixed an issue where relationships between sheets and fields were not being harvested under certain conditions, ensuring more complete lineage capture.

PostgresSQL collector: Now harvests all view definitions, regardless of user select privileges, ensuring more complete metadata capture.

OpenAPI collector:

Databricks collector:

dbt Core and dbt Cloud collectors: Now extract the host from the database-server option when it is provided as a URL, ensuring correct connection details.

Tableau collector:

Marquez collector: Added support for AWS S3, enabling metadata harvesting for S3 datasets through Marquez.

Oracle collector: Now harvests Oracle roles, extending coverage of user and access metadata.

Go here to access release notes for previous version.