Skip to main content
XenonStack

Big Data Engineering

XenonStack Blog Azure Data Catalog Retired: Purview Unified Catalog Guide

| Published | Updated

Azure Data Catalog Retired: Purview Unified Catalog Guide
20:23

Key takeaway

  • Azure Data Catalog retired on 15 May 2024. Learn how Microsoft Purview Unified Catalog now handles data discovery, glossary, lineage and access.

Azure Data Catalog, Microsoft's crowdsourced service for registering and discovering enterprise data sources, was retired on 15 May 2024. Microsoft now points customers to Microsoft Purview for data catalog features, and the current experience is Microsoft Purview Unified Catalog in the Microsoft Purview portal, built on top of the Purview Data Map. If your teams still talk about "the Azure Data Catalog", the practical question is no longer how to configure it but how to carry its job, helping people find, understand and request the right data, into Purview.

This guide is for data, platform and governance leaders who own that move. It covers what Azure Data Catalog did, how its concepts map to Unified Catalog, which sources the Data Map can scan, a sensible adoption sequence and the governance and cost questions to settle first. It stays on cataloguing and discovery; for data quality rules, scores and alerts in Purview, see our companion guide to Microsoft Purview data quality.

What was Azure Data Catalog, and what happened to it?

Azure Data Catalog was a fully managed cloud service that acted as a system of registration and a system of discovery for enterprise data sources. When a source was registered, the data stayed where it was; the catalog stored a copy of the source's structural metadata (such as column names and data types) and a reference to its location, then indexed that metadata for search. Any user could then enrich an entry with descriptions, tags, documentation and instructions for requesting access, which is why Microsoft described it as a crowdsourcing model for metadata.

Microsoft's own Azure Data Catalog documentation now opens with a retirement notice: Azure Data Catalog was retired on 15 May 2024, and Microsoft directs customers to Microsoft Purview for data catalog features. The underlying idea of a single, searchable inventory that both technical and business users contribute to did not go away. It moved into a broader governance product. If you are new to how catalogs are put together, our guide to data catalog architecture explains the core components.

What replaces Azure Data Catalog going into 2027?

The replacement is Microsoft Purview, and within it, Unified Catalog. Microsoft also renamed and reorganised Purview along the way, so it helps to be precise about which product name refers to what. The table below summarises the status of each product as stated on Microsoft Learn when we checked in November 2026.

Product What it is Current status What to do
Azure Data Catalog Original crowdsourced registration and discovery service Retired on 15 May 2024 Treat as legacy; move remaining knowledge into Purview
Microsoft Purview Data Catalog (classic), previously Azure Purview Classic governance portal at web.purview.azure.com with Data Map, catalog, glossary and Data Estate Insights No longer taking new customers; in customer support mode Existing accounts can be upgraded to the new portal experience
Microsoft Purview Unified Catalog Current data governance experience at purview.microsoft.com: governance domains, data products, glossary terms, OKRs, critical data elements, health management and data quality Current; Microsoft says new features are only available in the new portal Use for all new catalogue and discovery work
Microsoft Purview Data Map The metadata foundation that registers and scans sources, extracts schema, applies classifications and captures lineage Current; feeds Unified Catalog Register and scan sources here first

One naming point causes confusion in planning documents: "Azure Purview" became "Microsoft Purview", and the classic Purview catalog is now labelled "classic". When a vendor proposal or internal runbook mentions Azure Data Catalog or the Azure Purview studio, check whether it describes a retired or classic experience before you build on it.

How do Azure Data Catalog concepts map to Microsoft Purview?

Most of what teams did in Azure Data Catalog has a direct home in Purview, usually with more structure around ownership and access. The mapping below is the quickest way to brief data stewards who learned the old tool.

Diagram mapping retired Azure Data Catalog tasks (register, annotate, glossary, access notes) to Microsoft Purview (Data Map scans, governance domains, enterprise glossary, data products with access policies)
Azure Data Catalog Microsoft Purview equivalent What changes
Register a data source with the registration tool Register the source in Data Map, then scan it Scans extract technical metadata and can apply classifications; scans can be scheduled
Annotate assets with descriptions and tags Curate assets and attach glossary terms in Unified Catalog Curation is tied to owners and governance domains rather than open to everyone by default
Business glossary Enterprise glossary, under Unified Catalog > Discovery Glossary terms live inside governance domains and can carry access and data-handling policies
Search the catalog Data asset search and browse, under Unified Catalog > Discovery > Data assets Natural language search for data products is in preview
Document how to request access Data products with Unified Catalog access policies Self-service access requests with configurable approvals and attestations
Not available Lineage, classifications, data quality and health management Governance signals sit beside discovery in the same catalog

The biggest conceptual shift is the data product. Microsoft defines it as a business concept with a name, description, owners and a list of associated data assets, such as tables, files and Power BI reports, grouped under a use case. A governance domain can hold many data products, while each data product is managed by a single domain and can be discovered across domains. In practice this replaces the old pattern of searching a flat list of tables and hoping the description is current. Our guide to metadata as a product covers how to run catalogue metadata with that kind of ownership.

Why does the data discovery problem still matter?

Microsoft's description of the problem Azure Data Catalog set out to solve is still accurate. Data consumers often do not know a source exists, cannot connect without the location or connection string, cannot find the documentation, have to track down the responsible expert offline and still do not know how to request access. Data producers find that descriptions stored in the source are ignored by client tools, documentation drifts out of date and maintaining it is an ongoing cost. Discovery based on tribal knowledge does not scale.

Two things have raised the stakes since Azure Data Catalog launched. First, the number of sources has grown across Azure, other clouds and SaaS platforms, so a catalogue has to cover more than one vendor's estate. Second, AI agents and copilots now search for and act on enterprise data, not only analysts. Retrieval for those agents needs the same signals people need: what the data means, who owns it and who is allowed to use it. Permission-aware retrieval of that kind is what AgentSearch provides for enterprise AI search and retrieval.

A catalogue supplies metadata, lineage, a business glossary and data quality signals. It does not, on its own, enforce which decisions an agent may take, record who authorised them or keep a trace of why. In our work with enterprise data teams, that gap only becomes visible when agents are expected to reason, act and justify their actions. ElixirData Context OS keeps governed context, versioned policy and decision traces in one decision layer above AI agents, and it complements a catalogue such as Purview rather than replacing it.

Which data sources can Microsoft Purview Data Map scan?

Azure Data Catalog supported a long list of sources, but users often pointed to gaps, for example around Azure Data Lake Storage Gen2 and Snowflake. Both are now Data Map sources. Azure Data Lake Storage Gen1 itself was retired on 29 February 2024, so that part of the old discussion is closed. Purview Data Map publishes a source-by-source capability matrix. A selection is shown below; always check the live matrix for your sources, because capabilities differ by source and some are in preview.

Source Automatic classification Data lineage Notes
Azure Data Lake Storage Gen2 Yes Limited Full, incremental and scoped scans; sensitivity labels and live view supported
Azure Blob Storage Yes Limited Lineage also captured when used in Data Factory or Synapse pipelines
Azure SQL Database Yes Yes (preview) Also supports sensitivity labels and live view
Snowflake Yes Yes Supports sensitivity labels on Data Map assets
Oracle and Teradata Yes Yes Lineage from views and stored procedures
SAP HANA and SAP Business Warehouse No No Technical metadata only
Power BI and Microsoft Fabric No Yes Available through live view
Amazon S3 and HDFS Yes Amazon S3 limited; HDFS no Multicloud and on-premises file sources
MongoDB, Amazon Redshift and Salesforce No No Metadata extraction only; Redshift scanning is in preview

Two practical limits from the documentation are worth noting early: Data Map cannot currently scan an asset with "/", "\" or "#" in its name, and sources behind a firewall need a self-hosted integration runtime, which for some sources requires extra drivers or a JDK on the runtime machine. Teams comparing cloud catalogues can also read our overview of the AWS data catalog.

How should you move from Azure Data Catalog to Purview Unified Catalog?

We did not find a documented one-step import from Azure Data Catalog into Purview. Microsoft's guidance for moving between Purview accounts is to re-create configuration through the APIs and run fresh scans rather than copy assets, and the same logic applies here: scan the sources again and carry over the curated knowledge deliberately. A sequence that works for most enterprises:

  1. Inventory what the old catalogue held. List the registered sources, the descriptions and tags people relied on, the glossary and any access instructions. Most of the value is in the curated text, not the technical metadata.
  2. Start in the new portal. Use purview.microsoft.com. The free version of Purview data governance, currently in preview, lets you see metadata for supported Azure sources and search the catalogue before you commit; upgrading to the enterprise version carries your work across.
  3. Design governance domains and collections. Align domains with how the business works (functional units, lines of business, data domains, regulatory or project domains) and name owners for each.
  4. Register and scan sources in Data Map. Choose the integration runtime for each network zone, store credentials in Azure Key Vault and schedule scans so metadata stays current without manual updates.
  5. Rebuild the glossary as Unified Catalog terms. If you also used the classic Purview glossary, Microsoft provides a migration of classic terms into Unified Catalog (in preview). It is one-time and cannot be reversed, so run the sample migration first.
  6. Publish data products with access policies. Group assets by use case, attach owners and terms, and set approval rules so consumers can request access in the catalogue instead of chasing people offline.
  7. Add quality and health once discovery works. Data quality rules, scores and health actions build on the same domains and products; our Purview data quality guide covers that stage.

What governance, security and cost questions should leaders settle first?

A catalogue is part of your control environment, so a few decisions are best made before the first scan.

Question What Microsoft documents Why it matters
How many accounts can we have? One Microsoft Purview account per tenant, unless your organisation already had several under the earlier quota Test and production cannot simply be separate accounts; plan domains and collections instead
Who can see and curate what? Roles are managed at tenant level in the new portal; assets added through the REST APIs in the free version are visible only to data curators Map stewards, owners and readers to roles before inviting business users
What network changes are needed? The new portal uses purview.microsoft.com and api.purview-service.microsoft.com; Microsoft advises adding these to firewalls as legacy endpoints are retired Avoids broken access when classic endpoints go away
Are any access controls retiring? Purview access policies for SQL (DevOps, data owner and self-service policies) retire on 30 October 2027; Microsoft recommends SQL native roles and permissions for Microsoft Entra groups Do not design new SQL access around a feature with an end date
What does it cost? The free version is available to test; the enterprise version is paid, with pricing on the Microsoft Purview pricing page Scope scans and data quality runs to the domains that need them

Where AI agents will act on catalogued data, governance has to extend from who may read data to what an agent may do with it. AgentGRC keeps autonomous agent actions inside policy, with the audit trail risk teams need.

When is Purview Unified Catalog not the whole answer?

Purview is the natural choice for organisations whose estate centres on Azure, Microsoft Fabric and Power BI. It is less complete in some situations, and it is better to plan for them than to discover them mid-rollout:

  • Mixed estates with existing catalogues. If Databricks Unity Catalog or another catalogue already governs part of the estate, Purview can scan Unity Catalog metadata and lineage, but you still need to decide which system is the source of truth for each domain.
  • Sources with thin support. Several sources, such as MongoDB, Salesforce and SAP HANA, provide technical metadata without automatic classification or lineage, so stewards will curate more by hand.
  • Organisations without owners. Governance domains and data products only help if someone owns them. Without named owners, a new catalogue repeats the stale-documentation problem of the old one.
  • Decision governance for AI. A catalogue describes data. It does not decide what an agent is authorised to do or keep decision traces; that needs a separate context and policy layer.

To map one data discovery decision your agents should make, and the catalogue context and policy it depends on, book a Context Workshop. See what happens in the workshop.

What should you plan for in 2027?

One dated change in Microsoft's documentation affects catalogue access design.

  • 30 October 2027: Microsoft states that Purview access policies for SQL, covering DevOps, data owner and self service policies, retire on that date (Microsoft Learn, Purview access policies for SQL retirement).
  • After that date: SQL no longer enforces those policies and removes any access granted through them. Until then, existing policies keep working as they do today.
  • Before then: Microsoft recommends SQL native roles assigned to Microsoft Entra users and groups. Test in nonproduction and validate access before you remove the Purview dependency.

Do not design new SQL access around a feature with an end date.

Frequently asked questions

Is Azure Data Catalog still available?

No. Microsoft's documentation states that Azure Data Catalog was retired on 15 May 2024 and directs customers to Microsoft Purview for data catalog features.

What is the difference between Azure Data Catalog and Microsoft Purview Unified Catalog?

Azure Data Catalog was a crowdsourced registry for finding and annotating data sources. Unified Catalog adds governance domains, data products with access policies, glossary terms that carry policies, lineage, classifications, data quality and health management, built on scans from the Purview Data Map.

Is Azure Purview the same as Microsoft Purview?

Yes. Azure Purview was renamed Microsoft Purview. Its original catalog experience is now called Microsoft Purview Data Catalog (classic), which is no longer taking new customers and is in customer support mode, while new features go to the Microsoft Purview portal.

Can I import my Azure Data Catalog content into Purview automatically?

We did not find a documented direct import. Plan to register and scan your sources in Data Map and to carry over curated descriptions, tags and glossary terms deliberately. If you used the classic Purview glossary, Microsoft offers a one-time migration of classic terms into Unified Catalog, currently in preview.

Does Microsoft Purview support Snowflake and Azure Data Lake Storage Gen2?

Yes. Both are listed as Data Map sources with automatic classification and sensitivity labels; Snowflake also supports lineage, and Azure Data Lake Storage Gen2 has limited lineage plus lineage through Data Factory and Synapse pipelines. Azure Data Lake Storage Gen1, which Azure Data Catalog supported, was retired on 29 February 2024.

How many Microsoft Purview accounts can an organisation have?

Microsoft documents one Microsoft Purview account per tenant, unless the organisation already had multiple accounts under the earlier quota. Governance domains and collections are used to separate areas of the estate inside that account.

Follow XenonStack on LinkedIn: XenonStack on LinkedIn for new guides on agentic AI, data and platform engineering.

Ready to put AI agents to work? Explore ElixirClaw, the agentic OS platform to build, orchestrate and govern AI agents.

Table of Contents

dr-jagreet-gill

Dr. Jagreet Kaur

Chief AI Officer, XenonStack

LinkedIn profile

Dr. Jagreet Kaur specializing in Generative AI for synthetic data, Conversational AI, and Intelligent Document Processing. With a focus on responsible AI frameworks, compliance, and data governance, she drives innovation and transparency in AI implementation

Get the latest articles in your inbox

Subscribe Now

Interested in solving your challenges with the XenonStack team?
Get Started

Tell us about you and your priorities so we can prepare the right conversation. It takes about 2 minutes.

Please Select your Industry
Banking
Fintech
Payment Providers
Wealth Management
Discrete Manufacturing
Semiconductor
Machinery Manufacturing / Automation
Appliances / Electrical / Electronics
Elevator Manufacturing
Defense & Space Manufacturing
Computers & Electronics / Industrial Machinery
Motor Vehicle Manufacturing
Food and Beverages
Distillery & Wines
Beverages
Shipping
Logistics
Mobility (EV / Public Transport)
Energy & Utilities
Hospitality
Digital Gaming Platforms
SportsTech with AI
Public Safety - Explosives
Public Safety - Firefighting
Public Safety - Surveillance
Public Safety - Others
Media Platforms
City Operations
Airlines & Aviation
Defense Warfare & Drones
Robotics Engineering
Drones Manufacturing
AI Labs for Colleges
AI MSP / Quantum / AGI Institutes
Retail Apparel and Fashion

Continue