Trino

The Trino plugin discovers all catalogs (connected data sources like PostgreSQL, Hive, Iceberg, S3, etc.), their schemas, and tables/views.

Required Permissions

The connecting user needs SELECT access to system.metadata.catalogs, system.metadata.table_comments, and each catalog's information_schema. A read-only user with access to these system tables is sufficient.

AI Enrichment

When your Trino instance has AI functions configured, the plugin can automatically enrich discovered assets:

• Auto-generate descriptions (ai_generate_descriptions: true) — Uses the AI connector's ai_gen function to produce one-sentence descriptions for tables that have no comment.
• Auto-classify tables (ai_classify_tables: true) — Uses the AI connector's ai_classify function to assign a category label (e.g., analytics, pii, financial) to each table, added as a tag like ai-category:pii.

AI Setup

1. Configure an AI connector in your Trino installation (e.g., ai.properties)
2. Set ai_catalog to the catalog name of that connector
3. Enable ai_generate_descriptions and/or ai_classify_tables
4. Optionally customise ai_classify_labels and ai_max_enrichments

AI enrichment is best-effort - failures are logged as warnings but do not prevent normal discovery from completing.re logged as warnings but do not prevent normal discovery from completing.

name: my-trino-pipeline
runs:
  - trino:
      host: "<host>"
      user: "<user>"