Latest Release

The following changes have occurred to the open metadata types.

• Specific subtypes for PersonRole and GovernanceRole have been deprecated. PersonRole has a new attribute called actorRoleGroups to allow specific groupings to be defined. Egeria includes a number of predefined groups in the CoreContentPack as ValidMetadataValues.

  • TeamMember - person who is a member of a team.
  • TeamLeader - person who is a leader of a team.
  • ProjectManager - person leading a project.
  • NoteLogAuthor - person controbuting to a note log.
  • CommunityMember - person who is a member of a community.
  • CrowdSourceingContributor - person eho part of a crowdsourcing effort.
  • GovernanceOfficer - person leading a governance domain.
  • GovernanceRepresentative - Describes an advisor or representative from a governance domain that is working with other roles to help them with their governance tasks.
  • LocationOwner - Describes a role that has specific responsibilities with respect to a location.
  • BusinessOwner - Describes a role that has specific responsibilities with respect to a business area or service.
  • SolutionOwner - Describes a role that has specific responsibilities with respect to a solution.
  • AssetOwner - Role responsible for the correct management of an asset.
  • SubjectAreaOwner - person responsible for the definitions within a subject area.
  • ComponentOwner - person responsible for a component such as a software module.
  • DataItemOwner - person responsible for the correctness of a particular type of data value throughout its lifetime. Often these data values flow between systems and the DataItemOwner must be sure it is correct in all places.
  • DigitalProductManager - person responsible for the management and development of a digital product.

• New types called Perspective, Question, Skill, AssociatedSkills, and SkillSet have been added. This is to support context intelligence for AIs.

• New types called DataSharingHub and DataSharingRequest have been added. This is to support data sharing hubs.
• New collection type called WorkingSet. This is to support the new Resource Explorer AI application.
• New external reference type called ExternalStandard. This is to support links to external standards.
• New classification type called AccountingCodes. This is to support management of costs in projects such as system inventories.
• New governance engine and service types called ExplorerActionEngine and ExplorerActionService have been added. They describe the services that the new Resource Explorer application runs, and allow those requests to be exchanged with the Engine Host.
• The GovernanceStrategy entity type has been deprecated from GovernanceDriver. This means the governance drivers are more focused on the external factors that drive governance and the response from the governance teams is doumented in the governance policies.
• A new Note type has been added as a subtype of Notification, giving stewards a lightweight way to add an opinion or extra context to a note log.
• A new Risk entity type has been added, along with a new Regulator relationship linking external regulatory bodies to the governance drivers they enforce.
• New naming standards classifications allow naming conventions to be defined and checked for metadata elements.
• A legal attribute has been added to every entity type so that legal status and considerations can be recorded consistently across the metadata landscape.
• New types support the Smart Collections pattern, used by the new SmartCollections integration connector (see below).
• SolutionLinkingWire is now a multi-link relationship, and volumetric properties can now be recorded on data lineage relationships.
• A label attribute has been added to the ResourceList relationship.

The Liskov Data Sharing Hub Connector supports the management of a data sharing hub. Its main purpose is to derive a data dictionary for the data stores that are members of the data sharing hub.

Asset Maker API now supports the maintenance and querying of Software Capabilities.

Solution Architect API now supports the maintenance and querying of Design Patterns, including their nested, related and specialized relationships to other design patterns.

Privacy Officer API supports the maintenance of data processing descriptions associated with data processing purposes using in data privacy and data sharing agreements.

• Glossary Manager API now supports the maintenance of questions linked to glossary terms.
• Retrieving type definitions now allows the caller to control which attributes are returned, reducing the size of the response for callers that only need a subset of a type definition.
• Automated Curation API and Asset Maker API have new calls that allow the Resource Explorer to work with explorer action engines and the software capabilities that support them.
• The relationship calls in the Solution Architect API now honour the requested delete method, so relationships are removed in the same way as the elements they connect.
• Element headers now return an element's glossary classifications; previously the location classifications were returned in their place.

The PostgreSQL database connectors introduced in 6.0 have been joined by connectors for Oracle, MS SQL Server, DuckDB and (untested) Db2 LUW, all following the same cataloguing and survey pattern. MS SQL Server also gained a dedicated surveyor and cataloguer, and the shared JDBC Integration Connector that underpins these connectors has been refactored and hardened.

The JDBC resource connector that underpins the database connectors, the PostgreSQL repository and the PostgreSQL audit log destination now manages its database connections with a HikariCP connection pool, replacing the per-thread connection cache used previously. The pool is controlled by new configuration properties - jdbcMaximumPoolSize (default 10), jdbcMinimumIdle, jdbcConnectionWaitTimeout, jdbcMaximumConnectionLifetime, jdbcConnectionKeepAlive and jdbcConnectionLeakThreshold. A connector embedded in a repository or an audit log destination serves a whole server and wants the default pool size or larger, whereas a connector created for each catalog target is usually better set lower, since its pool size is multiplied by the number of catalog targets. The PostgreSQL repository connector and audit log destination accept jdbcMaximumPoolSize on their own configuration properties and pass it through to the connector they embed, so the pool can be sized where the server is configured rather than in the template that supplied the connection.

Callers of the connector must close the connections they take out - use try-with-resources - and commit their own writes, because a connection that returns to the pool with an open transaction is rolled back.

Connection and transaction leaks have also been fixed in the PostgreSQL, Oracle, MS SQL Server and Db2 LUW cataloguing and survey connectors, and in the PostgreSQL repository connector.

The Unity Catalog integration connector and content pack have been redesigned around namespace-based catalogs, with improved support for table columns and function parameters.

The new SmartCollections integration connector catalogues smart collections and keeps their membership up to date. See the Smart Collections documentation for more details.

The OpenAPI integration connector has been completed, and now extracts schema information from the APIs it catalogues. Its default refresh interval has been extended, since cataloguing an API specification does not need to be repeated often.

Egeria's default REST client connector is now a JDK-native HTTP client implementation, replacing the Spring-based SpringRESTClientConnector used previously.

The Open Metadata Framework (OMF) bean implementation has been completed, giving connectors and services a consistent way to work with open metadata beans.

• Governance actions can now be attached to entries in the digital library.
• Lineage is now captured for audit log connections to a topic.
• The information supply chain graph has been extended, and engine actions and notification subscribers can now be attached to supply chains.
• getAssetLineageGraph has been fixed, and governance action types can now declare producedGuards.

• Zone security has been added, giving finer control over which users can see which assets.
• Query results filtered out by security rules are now clearly distinguished from an explicit invalid-access error.
• getUserList has been added to the Security Officer and Privacy Officer APIs.
• The sample self-signed certificates shipped with Egeria have been regenerated with corrected validity dates and hostnames.
• The new JDK-native REST client connector now honours strict.ssl=false, which the previous Spring-based client picked up automatically. java.net.http.HttpClient ignores the JDK-wide defaults that setting relies on, so the connector now applies it directly when certificate and hostname checking has been switched off.

Mermaid graph titles are now protected against special characters that previously broke rendering, and reports can now include pie charts and mind maps alongside the existing graph types.

Egeria Workspaces has grown a lot of new capability since 6.0:

• The web portal gained an Insights search and analytics area, an Overview dashboard with business-value and AI-readiness tiles, and a Favorites feature, alongside improvements to the existing Catalog, Explorer, Audit and Operations tiles.
• Quickstart and Freshstart - plus new -multi-host variants for deployments that need to be reachable from other machines - gained a nanny integration daemon, a watchdog process, auto-deployed Unity Catalog containers, configurable memory limits, and automatic recovery from an out-of-memory container.
• A new set of optional associated runtimes lets you bring up Egeria alongside Airflow+Marquez, Superset, MLflow, Apache Atlas, Dagster, Prefect, Milvus and Delta Lake+Spark.
• Dr.Egeria commands can now be driven from Obsidian via a new MCP/SSE-based plugin, with session locking so a shared vault can be used safely by a team.

• Dr.Egeria has had a substantial rewrite: a two-pass parser now supports forward references within a batch of commands, and new command families cover Action Author (governance action processes, types and steps), Design Patterns, Risk, DataSharingHub/DataSharingRequest, and Perspective/Question/Skill/SkillSet.
• A new Report family adds Dashboard Sheet authoring and a set of reusable analytic functions (business-value signals, ownership coverage, metric trend, term-definition completeness, active contributors) with Vega-based chart generation, underpinning the Egeria Workspaces portal's new Insights and Overview dashboards.
• My Egeria can now run as a browser app as well as a terminal UI, and gained journal and blog entries, and management of your to-dos, meetings, reviews, roles, communities and teams.
• pyegeria parameters have been harmonized across the view service clients, with broader use of kwargs, typed exceptions with automatic token renewal, and new Privacy Officer and Connection Maker modules to match the equivalent Java-side services.
• The MCP server has been migrated to the mcp 2.0.0 MCPServer API.

The Coco Pharmaceuticals content pack has been refreshed several times, adding a new sustainability glossary and scenario, new information supply chain and governance definition content, new reference data sets, and new tabular data sets for licenses, exceptions and certifications. A Manufacturing governance domain, led by Stew Faster, has been added to the governance program content, along with its governance officer and community.

Two new test suites have been added to the build. The build verification test (BVT) suite exercises the platform's Java clients - assets, collections, projects, platform and server operations - against a running OMAG Server Platform, and the query FVT suite covers paging, sorting, status handling and subtype filtering. Both start and stop the platform themselves. A number of fixes to the PostgreSQL repository came out of writing the query FVT suite.