Skip to main content

Enabling Ontology & Knowledge Graph

Ontology and the knowledge graph are two different things with two different requirements. Knowing which is which saves a lot of unnecessary infrastructure.

What Works with No Extra Infrastructure

The ontology is stored relationally. Everything below works on a stock OpenMetadata deployment:
  • The Ontology Explorer — graph, filters, and the Model and Data views
  • Ontology authoring through the API — concepts, typed relations, attributes, constraints, mappings, and releases
  • Relationship type management
  • Ontology Library pack installation
  • Ontology import and export (OWL, Turtle, N-Triples, JSON-LD, SKOS, CSV)
  • Glossary-scoped SPARQL (POST /api/v1/glossaries/{id}/sparql)
  • The ontology_describe MCP tool

What Needs the RDF Triple Store

  • The standalone SPARQL Playground
  • The Knowledge Graph tab on tables
  • Catalog-wide SPARQL, graph exploration, and graph export
  • Inference and SHACL validation over the catalog
  • Graph insights — importance, centrality, communities, paths, recommendations, tag analytics
  • The sparql_query, entity_neighborhood, find_by_tag, and shacl_validate MCP tools
  • Linked Open Data IRI dereferencing

What Needs an LLM Provider

  • The /api/v1/ontology/ai relationship, mapping, SPARQL, and domain-draft proposals
Requires RDF_AI_ENABLED=true (the older name RDF_ASK_COLLATE_ENABLED still works) and LLM features turned on (LLM_ENABLED=true) with a provider set in LLM_PROVIDER (openai, azureOpenAI, bedrock, google, or anthropic).

Turning on RDF

The RDF knowledge graph is Beta and disabled by default. Turning it on for a large existing catalog adds meaningful load while the initial index runs. Test in a non-production environment first and schedule the initial index for a low-traffic window.
Full deployment instructions — Docker, Kubernetes, Helm, sizing, and troubleshooting — are in Enable RDF (Knowledge Graph). The short version:
1

Run Apache Jena Fuseki

Reachable from the OpenMetadata server. The repository ships one at docker/rdf-store.
2

Set the environment variables on both containers

The migrations container loads the ontology and shapes graphs and registers RdfIndexApp; the server container does everything else. If only one has RDF configured, installation fails with RdfRepository not initialized.
3

Run the initial index

Settings → Applications → RDF Knowledge Graph Indexing → Run Now, or:
4

Verify

Expect enabled: true and projectionState: READY. REBUILDING means the index is still running; DEGRADED means something needs attention.

Trying It Locally

The repository ships a startup script that brings up the whole stack with RDF and Fuseki enabled:
It sets sensible development defaults, including materialized inference.

Configuration Reference

All settings live under rdf: in openmetadata.yaml and are overridable by environment variable.

Core

Connection and Write Tuning

Inference

RDF_INFERENCE_ENABLED, RDF_DEFAULT_INFERENCE_LEVEL, RDF_MAX_IN_MEMORY_INFERENCE_TRIPLES, and RDF_CACHE_INFERRED_TRIPLES are deprecated. The server ignores them. The in-process reasoner they controlled was removed in 2.1, so rdfs and owl inference requests fail with 503. See Reasoning & Validation.

Validation and Profile

Linked Data, AI, Federation

Federation is a data-exfiltration surface. A SERVICE clause sends parts of your query — and potentially your data — to a third-party endpoint. It is disabled by default, and the allowlist is empty. Enable it only for endpoints you control or explicitly trust.

Indexing Application

RdfIndexApp configuration (Settings → Applications): Only one reindex job can be active per cluster at a time.

Inference Application

RDF Inference Materialization (RdfInferenceApp) is the second application RDF registers. It materializes enabled inference rules into their durable named graphs on a schedule — every 5 minutes by default — and only rebuilds rules whose dirty flag is set. Select Run Now to materialize immediately. It only does useful work when RDF_MATERIALIZED_INFERENCE_ENABLED=true. See Reasoning & Validation.

Verifying the Whole Thing

In the UI, the SPARQL Playground at /governance/sparql should run queries, and any table should show a Knowledge Graph tab.

Next

Deployment guide

Docker, Kubernetes, Helm, sizing, and troubleshooting.

Knowledge Graph

What you get once it is on.