Skip to main content

Unity Catalog Business Glossary and Domains: GA or Preview?

Emily Winks, Data Governance Expert, Atlan
Data Governance Expert
Updated:
|
Published:
11 min read

Key takeaways

  • Domains moved through all three stages in under eight months.
  • Every Page belongs to exactly one domain or subdomain, so you set domains up first.
  • Public Preview is supported for production with an SLA.
  • Accepting one batch of suggested edits clears every other pending batch on that Page, and that cannot be undone.

Unity Catalog business glossary and domains: which features are GA and which are preview?

Unity Catalog domains and subdomains are generally available (GA), together with the Discover page, since 8 October 2026. The business glossary ships as Pages, which is in Beta, a stage the Azure Databricks release-type table marks as not for production. Genie Ontology, which reads both, is in Public Preview. Releases are staged, so the release notes warn that an account can lag the release date by a week or more.

Is your AI agent context production-ready?


Unity Catalog domains and subdomains are generally available (GA), together with the Discover page, since 8 October 2026. The business glossary ships as Pages, which is in Beta, a stage the Azure Databricks release-type table marks as not for production. Genie Ontology, which reads both, is in Public Preview.

Map Your Unity Catalog Coverage


Give it every system in your estate and which catalog governs each today. It returns a coverage table and the specific gap ranked by what breaks first. Read the skill.

Paste into a new chat

Use the skill at https://atlan.com/skills/catalog-coverage-gap-map.md to map catalog coverage across our Unity Catalog estate. Ask me for whatever it needs.

Run once in a terminal

curl -fsSL --create-dirs \
  -o ~/.agents/skills/catalog-coverage-gap-map/SKILL.md \
  https://atlan.com/skills/catalog-coverage-gap-map.md

For an agent

curl -fsSL https://atlan.com/skills/catalog-coverage-gap-map.md

Which Unity Catalog business glossary and domains features are GA, and which are preview?

Every status below is the one the linked Azure Databricks page states on 28 October 2026, the date we checked each link.

Feature Status Since Source
Domains and subdomains GA 8 Oct 2026 October 2026 release notes
Discover page GA 8 Oct 2026 October 2026 release notes
Domains REST API Beta not dated Domains and subdomains
Account-scoped Discover Beta, on by default not dated Discover page
Pages (business glossary) Beta 12 Aug 2026 August 2026 release notes
Genie Ontology Public Preview not dated Genie Ontology
Certified and deprecated tags No preview label not dated Flag data as certified or deprecated
Tag automations Beta 7 Aug 2026 August 2026 release notes
Tag propagation Beta 12 Oct 2026 October 2026 release notes
Metric views No preview label not dated Unity Catalog metric views
Metric view window measures GA 6 Oct 2026 October 2026 release notes
Genie-recommended domains No preview label not dated Discover page
Governance Hub (Tags page has a Domains tile) Beta not dated Governance Hub

Read two rows first. Domains are production-ready. The glossary is not: the Pages documentation still marks it Beta. Releases are staged, so the release notes warn that an account can lag the release date by a week or more.


What does Beta mean for Pages, compared with GA for domains?

The labels carry real support terms. The Azure Databricks release-type table defines them like this:

Stage For production Interface stable SLA Support
Beta No No No Engineering team
Public Preview Yes Yes Yes Support team
GA Yes Yes Yes Support team

Admins control access to Beta features. For Pages, account admins do that from the account console Previews page, as the Pages documentation states. Beta features are documented and intended to move to Public Preview or GA, and a notice appears in the UI when GA is close.

In practice, a team can build its domain model on GA footing today. Glossary content in Pages is a different decision. The interface can still change, nothing carries an SLA, and support comes from the engineering team instead of the support team. If a production Genie One rollout depends on a definition, keep that definition somewhere you can export it from.

Genie Ontology sits in between. Public Preview is supported for production with an SLA. The Genie Ontology page also notes that curating ontology snippets costs nothing at this time, and that pricing may change after notice.


How did domains and the business glossary reach their current status?

Domains moved through all three stages in under eight months. The glossary was announced first under a different name.

This history explains why answers disagree. Flexera’s June recap, for example, lists domains in Public Preview and the glossary as “Preview coming soon”. Both were true when they were written. Search the docs for “Pages” as well as “glossary”, because the shipped feature uses the new name.


What can you set up with Unity Catalog domains today?

A domain is an account-level grouping built on a governed tag. Per the domains documentation, you add an asset to a domain by adding its tag. Tag a catalog or schema and the tables inside inherit the domain.

The rules that shape your design:

  • One level of nesting: a subdomain such as Finance/Tax cannot hold its own subdomains.
  • Many-to-many membership: an asset can sit in several domains, and subdomain assets still appear in the parent domain.
  • Separate tags: tagging an asset Finance/Tax does not also tag it Finance.
  • Curator rights: you need MANAGE DISCOVER at the account level, or MANAGE on one domain or subdomain. Account and workspace admins hold MANAGE DISCOVER by default.
  • Draft first: a new domain stays visible only to curators until someone publishes it.

Limits to plan around: subdomains are managed only from a workspace-level Manage Discover page, and the REST API and the databricks_domain Terraform resource cannot customize Discover sections. On workspaces with the compliance security profile, a workspace admin must turn the Discover Page on. Since domains are governed tags, the governed tags caps apply too: 1,000 governed tags per account and 500 allowed values per tag.

Domain names, descriptions, and subtitles are replicated globally, so the docs warn against putting personal or sensitive information in them.


What do Pages hold, and what should you watch while they are in Beta?

A Page is a governed definition of one business concept: a term, an entity or an acronym. The Pages documentation lists its fields as domain, owner, synonyms, description, a rich-text body, related assets, and sources. Every Page belongs to exactly one domain or subdomain, so you set domains up first.

Permissions come from the domain. Anyone with access to the domain can read its published Pages, and consumers can create, publish, comment, and suggest edits. Page owners and curators edit directly. Genie Code can draft one Page from source material, or bulk import a set of terms from files, MCP connections, and public links, flagging conflicts and duplicates for review.

The payoff is in Genie One. When a question touches a concept defined in a Page, Genie One prioritizes the Page over inferred context and cites it. Draft Pages reach only the owner’s own Genie One conversations.

Four details are worth knowing before you load hundreds of terms:

  1. Page data uses Microsoft-managed keys; customer-managed keys are not supported “at this time”.
  2. Pages are account-level, so content may be replicated globally. Keep regulated data out of names, synonyms, and body text.
  3. Accepting one batch of suggested edits clears every other pending batch on that Page, and that cannot be undone.
  4. Activity is logged in system.access.audit under the pages service.

How do metric views, certification and Genie Ontology fit with the glossary and domains?

Databricks groups four components as Unity Catalog semantics: metric views for KPIs, domains for business areas, Pages for definitions, and certification for trust. Together they form the human-modeled layer of Genie Ontology.

The Genie Ontology adds a second layer of inferred snippets drawn from metric views, dashboards, SQL queries, and Genie Agents. Each snippet gets an authority score from its source, usage, and freshness, and Unity Catalog permissions decide which snippets a user’s answer can draw on.

So one Genie One answer can rest on a GA domain, a Beta Page and a Public Preview ontology at the same time. That is the reason to track status per feature rather than per product.

Certification is the cheapest of the four to adopt. It is the system tag system.certification_status, with the values certified and deprecated, and it applies to tables, views, dashboards, Genie Agents, notebooks, and more. The certification documentation carries no preview label, though automating it at scale relies on tag automations, which are in Beta. For how governed KPIs relate to a wider context layer, see this guide to Unity Catalog metrics.


How does Atlan work with Unity Catalog domains, tags and metric views?

Atlan connects through governed tags, the primitive domains and certification are built on, and through metric views. Atlan imports Unity Catalog tags and matches them to Atlan tags. With reverse sync on, a tag updated in Atlan is updated on the Databricks asset; reverse sync changes existing tags and never creates or deletes them. Atlan’s crawler maps Databricks metric views to a DatabricksMetricView asset type, though upstream column lineage for dimensions and measures is not supported.

Context can also flow the other way. Deploying from Atlan’s Context Engineering Studio to Databricks Genie pushes descriptions into Unity Catalog, creates one metric view per table in the target schema, and configures a Genie Space on those views.

General Motors runs both. As Brian Ames of General Motors put it on Atlan’s Databricks partner page: “If we didn’t have AI in our arsenal, we could find ourselves at a competitive disadvantage. Unity Catalog worked out of the box for us… and Atlan gave us visibility from the cloud all the way back to our on-prem.”

For the glossary side, compare how a business glossary differs from a data catalog, and how teams manage KPI definitions and a glossary in a catalog.


FAQs about Unity Catalog business glossary and domains

Is the Unity Catalog business glossary generally available?


No. On 28 October 2026 the glossary ships as Pages, which the Pages documentation marks as Beta. It entered Beta on 12 August 2026. Beta features are not intended for production, carry no SLA, and are supported by the engineering team. Account admins enable Pages from the account console Previews page.

Are Unity Catalog domains GA on Azure Databricks?


Yes. The October 2026 release notes list the Discover page and domains as generally available from 8 October 2026. The GA note covers subdomains too. Managing domains through the REST API is still Beta, and releases are staged, so an account can see the change a week or more later.

Do I need domains before I can create Pages?


Yes. Every Page is attached to exactly one domain or subdomain, so the Pages documentation asks for at least one domain first. A Page then inherits read access from its domain. Curators with MANAGE DISCOVER or MANAGE on the domain can edit and publish any Page in it.

Can I manage Unity Catalog domains as code?


Partly. The domains documentation describes a domains REST API in Beta and a databricks_domain Terraform resource. Neither can customize Discover page or domain-page sections, which stay UI-only, and subdomains are managed from a workspace-level Manage Discover page.

Does Genie One use Pages and domains when it answers?


Yes. Both feed the Genie Ontology, which is in Public Preview. When a question matches a Page, Genie One prioritizes that definition over inferred context and cites the Page. Snippets are gated by Unity Catalog permissions, so each user’s answer draws only on sources that user can see.


Sources

  1. October 2026 release notes, Azure Databricks, Microsoft Learn
  2. August 2026 release notes, Azure Databricks, Microsoft Learn
  3. June 2026 release notes, Azure Databricks, Microsoft Learn
  4. February 2026 release notes, Azure Databricks, Microsoft Learn
  5. Azure Databricks preview releases, Microsoft Learn
  6. Pages, Azure Databricks, Microsoft Learn
  7. Domains and subdomains, Azure Databricks, Microsoft Learn
  8. Discover page, Azure Databricks, Microsoft Learn
  9. Genie Ontology, Azure Databricks, Microsoft Learn
  10. Unity Catalog semantics, Azure Databricks, Microsoft Learn
  11. Unity Catalog metric views, Azure Databricks, Microsoft Learn
  12. Flag data as certified or deprecated, Azure Databricks, Microsoft Learn
  13. Governed tags, Azure Databricks, Microsoft Learn
  14. Tags overview, Azure Databricks, Microsoft Learn
  15. Databricks Data + AI Summit 2026: All 20+ major launches, Flexera
  16. Manage Databricks tags, Atlan Documentation
  17. What does Atlan crawl from Databricks?, Atlan Documentation
  18. Deploy to Databricks Genie, Atlan Documentation

Share this article

signoff-panel-logo

Atlan is the Context Layer for AI. It translates business knowledge, including data definitions, working procedures, and governance policies, into context AI can actually use. This knowledge lives in a single Enterprise Data Graph that every team and AI agent can reach.

In Atlan's AI Labs benchmark, adding this context improved AI's text-to-SQL accuracy by 38%.

Atlan is recognized as a Leader across multiple Gartner reports and Forrester Waves, and is trusted by over 400 enterprises representing $10T+ in market cap, including Mastercard, Workday, General Motors, CME Group, HubSpot, FOX, Virgin Media O2, and Elastic.

Bridge the context gap.
Ship AI that works.