Alation Connector for Dataiku OCF

The Dataiku OCF Connector, developed by Logan Data, enables Alation users to automatically extract and catalog metadata from the Dataiku Data Science Studio (DSS) platform, bringing Projects, Datasets, Recipes, Models, Dashboards, Insights, Notebooks, Webapps, Analyses, and Charts into Alation's unified data intelligence platform.

Dataiku DSS is a collaborative data science platform used by organisations to build, deploy, and monitor machine learning pipelines and advanced analytics workflows. By connecting Dataiku to Alation through this OCF connector, data teams gain full visibility into how data flows from source systems into Dataiku transformations and ultimately into business insights, all searchable, trusted, and governed from a single catalog.

Built on Alation's Open Connector Framework (OCF) and delivered as a Zip file installable via Alation Connector Manager, this connector supports both Alation On-Premise and Alation Cloud Service deployments, with Alation Agent support from connector version 1.1.2.4595 for network-isolated environments.

Object mapping - what gets extracted and how it appears in Alation:

• Projects mapped to Alation Folders, capturing owner, creation time, modification time, project description, and project-level permissions
• Datasets mapped to Alation Data Sources, with all dataset columns cataloged as Data Source columns including name, datatype, and schema information
• Recipes mapped to Alation Reports, with recipe source code captured for lineage and transformation documentation
• Models mapped to Alation Reports, capturing model summary metadata for ML asset governance
• Dashboards mapped to Alation Folders, with owner, creation time, and modification time
• Insights mapped to Alation Reports, with report objects cataloged as measures and dimensions
• Notebooks mapped to Alation Reports, with notebook source code preserved for reference
• Webapps mapped to Alation Reports
• Analyses mapped to Alation Folders, with owner and timestamp metadata
• Charts mapped to Alation Reports, with chart objects cataloged as measures and dimensions

Lineage and governance:

• Cross-system lineage: map Dataiku datasets back to their origin data sources already cataloged in Alation, enabling end-to-end visibility from raw data through Dataiku ML pipelines to downstream business outputs
• Automatic lineage generation is enabled by default and can be disabled on a per-source basis for teams who prefer to manage lineage manually or via API
• Obfuscate Literals: hide sensitive literal values from ingested queries displayed in Alation's Queries tab, configurable per data source

Extraction and scheduling:

• Default metadata extraction (MDE) based on built-in connector queries
• Selective extraction with project-level include/exclude filters
• Automated extraction scheduling from within Alation, configure the day and time for recurring catalog updates without external orchestration
• Manual on-demand extraction available at any time from the Settings page

Authentication and connectivity:

• Basic authentication (username and password) using a dedicated Dataiku service account
• Read Project Content permission required at the project level
• Read Dashboards permission required at the global or project level
• Configurable logging levels (Log4j framework: DEBUG, INFO, WARN, ERROR) accessible from Admin Settings > Manage Connectors in Alation

Compose support: users can run queries against Dataiku datasets directly from Alation's Compose interface. Compose with Alation Agent is supported from connector version 1.1.2.4595.

Compatibility: Alation 2024.1.4 and above. Supported on Alation On-Premise and Alation Cloud Service with or without Alation Agent.