> ## Documentation Index
> Fetch the complete documentation index at: https://docs.open-metadata.org/llms.txt
> Use this file to discover all available pages before exploring further.

# 2.1 Release | OpenMetadata

> Release notes for OpenMetadata 2.1, covering new connectors, metrics, data quality, lineage, governance, and API features, with a link to upgrade your instance.

# 2.1 Release Notes

> [card] [Upgrade OpenMetadata](https://docs.open-metadata.org/v2.1.x-SNAPSHOT/deployment/upgrade.md)
>
> Learn how to upgrade your OpenMetadata instance to 2.1.0.

**2.1.0 Release**

November 20, 2026

For more information, see the [OpenMetadata 2.1.0 GitHub release](https://github.com/open-metadata/OpenMetadata/releases/tag/2.1.0-release).

OpenMetadata 2.1 adds richer Data Quality and governance, a larger Context Center, change requests for reviewing governance changes before they apply, and a complete Metrics entity. It also adds Ontology Explorer and a knowledge graph for your business vocabulary, seven new connectors, and many improvements to existing ones.

> [!WARNING]
>
> Before you upgrade, take a full database backup and review the [breaking changes](https://docs.open-metadata.org/v2.1.x-SNAPSHOT/deployment/upgrade/breaking-changes/overview.md). OpenMetadata 2.1 removes the old Feed API, retires the Delta Lake connector, and tightens permission checks on search and lineage.

## Features

The following sections describe what is new in this release.

### Data Quality

This release gives you more control over how tests fail and clearer insight into test runs.

* **Failure thresholds**: Decide how much failure a test tolerates before it fails. Set the threshold as a number of rows or as a percentage. Percentages must be between 0 and 100.
* **Data Quality Dimensions**: Group your tests by dimension, such as completeness or accuracy, to see data quality from the angle your team cares about.
* **Filter by more than one status**: In the Test Cases tab and in the API, filter test cases by several statuses at once, such as Failed and Aborted.
* **Clearer run history**: The result history card shows a date range and summary tiles for your runs. Each result now records how long the run took and any error. A new run-details card shows why a run aborted.
* **Incident Manager**: A new Grouped Incidents view puts related incidents together. The Open card now counts incidents that are acknowledged or assigned, and the cards are labeled as daily counts.

### Governance

These changes give you more control over workflows, policies, and how governed assets are organized.

* **Smarter workflows**: A workflow condition can check how many assets or output ports a Data Product has. A workflow can also start when a column's custom property changes.
* **Governance by service**: Tags on a service now apply to its databases, schemas, and tables. Policy conditions can use an asset's service, for example to hide every asset from one service.
* **Impersonation in policies**: New `isImpersonated()` and `impersonatedBy()` conditions let a policy react when a user is acting on behalf of someone else.
* **Lifecycle stages**: Define lifecycle stages for your entities.
* **Domain styles**: Give each domain its own color and icon.
* **Glossary settings**: New glossary settings, including management of relation types.
* **Data Retention**: A new `workflowRetentionPeriod` setting lets the Data Retention app clean up old automation workflows.

### Context Center

Context Center now makes it easier to find, preview, and share content.

* **Articles toolbar**: Filter and sort articles from a new toolbar.
* **File preview**: Open Markdown, text, PDF, and image files without leaving OpenMetadata.
* **File visibility**: Control who can see a file.
* **Tabular profiles**: See tabular profiles on assets.
* **Query runner**: Attach files to query runner requests.
* **Linked articles**: Link Context Center articles from entity pages.

### Change Requests

Change requests let you review governance changes before they are applied, instead of changing the catalog directly.

* Reviewers see a preview of exactly what will change, then approve or reject it.
* The person who raised a request can withdraw it.
* Admins can override a request when needed.
* Workflow runs now show how each approval was decided, so you can see who approved a change and why.

### Metrics

Metrics are now a complete, browsable part of your catalog.

* **Hierarchy and details**: See a metric's full hierarchy and a detail page with Dimensions and Measures on the overview and an Assets tab that lists the assets it relates to.
* **Metric Groups**: Organize related metrics together in a new Metric Group.
* **Column-level links**: Link a metric to a specific table column, not just the table. Column-level lineage between them is kept.
* **Keeps your edits**: When ingestion updates a metric, its hierarchy, reviewers, and experts are no longer lost.
* **More sources**: Unity Catalog metric views and Looker measures are now ingested as Metrics.

### Lineage

Hierarchical lineage maps let you start from a high-level view and drill down from service to database, schema, asset, and column. Large lineage graphs load one section at a time, and breadcrumbs always show where you are. Open the map from **Lineage** in the left sidebar.

### Alerts

Alerts are easier to set up and to monitor.

* **Alert health**: See whether an alert and its destinations are working.
* **Available options**: When you set up an alert, see which sources and conditions you can use.
* **Safer webhook secrets**: Webhook secrets are now encrypted with the server's Fernet key, the same way other secrets already are.

### Conversations and Announcements

Conversations have been rebuilt. They now support replies and filters, and each conversation records where it came from. Announcements now have a type, either Information, Warning, or Issue, so readers can quickly see how important an announcement is.

### Ontology Explorer and Knowledge Graph

Ontology Explorer is a read-only graph of your glossary terms and how they relate to each other. It is a Beta feature. The ontology APIs and the knowledge graph are also included in this release.

* **Explore your vocabulary**: See the terms and relations of every glossary in one graph, search for concepts, and switch between the Model and Data layers.
* **Ontology APIs**: Manage ontologies, concepts, properties, relationships, constraints, mappings, and releases, and run SPARQL Protocol and RDF Query Language (SPARQL) queries over an ontology's draft or active release.
* **Ontology Library and relationship types**: Install ready-made ontology packs, import and export ontologies, and manage relationship types.
* **Knowledge graph**: Turn on Resource Description Framework (RDF), which is off by default, to mirror every entity in your catalog into a triple store. This adds the Knowledge Graph tab, the SPARQL Playground, inference, validation, and the knowledge graph tools for AI agents.

### New Connectors

This release adds seven connectors.

* **Rill** (dashboard): Ingests Rill dashboards and their data models.
* **NATS JetStream** (messaging): Discovers streams and subjects, extracts schemas, and samples messages. Supports Basic, token, NKey, Transport Layer Security (TLS), and anonymous authentication.
* **ClickZetta** (database): Ingests metadata, usage and query lineage, and supports profiling, data quality, data diff, and dbt.
* **Salesforce Data360** (database and pipeline): Ingests Salesforce Data360 (formerly Data Cloud) as a database source and as a pipeline source.
* **Tableau** (pipeline): Ingests Tableau Prep flows and extract refreshes as pipelines.
* **Databend** (database): Ingests databases, tables, views, columns, and comments.
* **Neo4j** (database): Ingests graph node labels and relationship types as tables.

### Connector Improvements

These existing connectors and ingestion features have new options.

* **Microsoft Fabric**: Sign in with a certificate or a client secret.
* **SSH File Transfer Protocol (SFTP)**: Sign in with a username and password or with a key.
* **Druid**: Connect to brokers secured with TLS by using the `druid+https` scheme.
* **Hive**: Partition key columns are now marked in the metadata, so you can tell them apart from other columns.
* **Dagster**: Run status checks now honor the `statusLookbackDays` setting, which controls how many days back OpenMetadata checks run statuses.
* **dbt**: Assets are added to Data Products automatically when the dbt manifest sets `meta.openmetadata.dataProducts`.
* **Data contracts**: Importing an Open Data Contract Standard (ODCS) data contract now gives you an import report.
* **Run for a single asset**: Run an ingestion pipeline for one asset instead of the whole service.
* **Ingestion pipelines**: Set owners on a pipeline. Admins can force delete a pipeline when its runner is unavailable.

### Settings and Sign-in

These changes help you configure and test your deployment.

* **Test your login**: Test a single sign-on setup without signing out.
* **Where a setting comes from**: Settings show whether a value comes from the configuration file or from the UI.
* **Content delivery network (CDN)**: A new setting lets you serve UI assets from a CDN.
* **Applications**: The whole application card is clickable in **Settings** > **Applications**.

### API

The following API changes are included in this release.

* `GET /v1/services/overview` returns the number of services for each type, plus one sorted page of services, in a single call.
* New endpoints are available for ontologies, change requests, Metric Groups, Data Quality Dimensions, conversations, and RDF and SPARQL queries.
* The Model Context Protocol (MCP) `get_entity_lineage` tool supports column-level lineage and cursor paging.

## Deprecations

The following items are deprecated in OpenMetadata 2.1. Move to the replacement soon. Items marked as still working keep working in this release.

| Deprecated | Use instead | Status |
| - | - | - |
| Airflow as the internal orchestrator | The native Kubernetes Orchestrator | Still works in 2.1. It will be removed in a future release. |
| Classic mode of the interface | AI mode | Still works in 2.1. It will be removed in a future release. |
| Sign-in claim for user names (`jwtPrincipalClaims`) | The email claim (`emailClaim`) | Still works in 2.1. It is no longer a required field. |
| Sign-in claim mapping (`jwtPrincipalClaimsMapping`) | The email claim (`emailClaim`) and the display name claim (`displayNameClaim`) | Still works in 2.1. |
| Administrator names (`adminPrincipals`) | Administrator emails (`adminEmails`) | Still works in 2.1. It is no longer a required field. |
| Principal domain (`principalDomain`) | `botDomain` for bots, and `allowedEmailDomains` for user restrictions | Still works in 2.1. It is no longer a required field. |
| Allowed domains (`allowedDomains`) and `enforcePrincipalDomain` | `allowedEmailDomains` | Still works in 2.1. `enforcePrincipalDomain` is no longer a required field. |
| Secure socket connection (`enableSecureSocketConnection`) | Nothing to replace. Socket connections are always authenticated | The setting has no effect. You can remove it. |
| Alert resources endpoint (`/{alertType}/resources`) | `POST /capabilities` | Still works in 2.1. |

## Removed

The following connector is no longer supported in OpenMetadata 2.1.

| Removed | What happens on upgrade |
| - | - |
| Delta Lake connector | Existing Delta Lake services become Custom Database services, and their ingestion pipelines are deleted. |

[View the full changelog](https://github.com/open-metadata/OpenMetadata/compare/2.0.8-release...2.1.0-release)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.