Overview
Collate uses RDF (Resource Description Framework) to build a knowledge graph of your metadata. This enables semantic search, SPARQL (SPARQL Protocol and RDF Query Language) queries, and JSON-LD representations of your data assets.What RDF Indexing Enables
RDF indexing extends your metadata catalog with graph capabilities, making it possible to query and explore data relationships beyond simple search.- Semantic search — Query metadata using meaning and relationships, not just keywords.
- Knowledge graph — Explore connections between entities across your data catalog.
- SPARQL queries — Run graph queries directly against your metadata.
- JSON-LD serialization — Export any entity in standard linked data formats (JSON-LD, Turtle, RDF/XML, N-Triples).
- Ontology support — Metadata is modeled against the OpenMetadata OWL ontology.
How It Works
RDF indexing runs as a background application alongside your Collate deployment. Collate uses Apache Jena Fuseki as the underlying triple store. When RDF is enabled:- New and updated entities are automatically indexed to the triple store.
- Hard-deleted entities are removed from the triple store.
- Existing entities can be reindexed on demand via the RDF Index App.
Storage Maintenance
Starting in 2.0.8, scheduled maintenance and low-space protection compact the Fuseki store before its volume fills. During compaction, RDF writes pause safely and RDF write requests receive a maintenance response. Reads and the relational metadata catalog remain available. Storage status reports whether maintenance is active and how much free space remains. Keep the configured free-space reserve large enough for compaction to finish, and alert on the Fuseki persistent volume. If low-space maintenance starts repeatedly, expand the volume instead of lowering the reserve until the store has no working space.Available API Endpoints
Use these endpoints to check RDF status, retrieve entities in linked data formats, or run custom graph queries. Once RDF is enabled, these endpoints are available on your Collate instance:Supported RDF Formats
The default format is JSON-LD. Pass?format= to the entity endpoint to select a different serialization: