Databricks
Databricks integration on indie.money — run SQL, manage Unity Catalog, query vector search indexes, and serve models
Written By Philippe
Last updated 30 days ago
Databricks
✅ Supported — Run SQL statements, manage Unity Catalog objects, work with workspace files, query vector search indexes, and call model serving endpoints
Credential Setup
When configuring the Databricks credential in Secure Vault, provide:
- Host — Your Databricks workspace URL, e.g.
https://adb-1234567890123456.7.azure.databricks.com(the address you use to open the workspace in a browser) - Personal Access Token — A workspace access token, generated under Settings → Developer → Access tokens in Databricks. Starts with
dapi.
Operations
- Databricks SQL — Execute Query
- Unity Catalog — Create, Get, List, Update, and Delete for catalogs, tables, functions, and volumes
- Files — Upload File, Download File, Get File Info, Delete File, List Directory, Create Directory, Delete Directory
- Genie — Get Space, Start Conversation, Create Message, Get Message, Execute Message Query, Get Query Results
- Model Serving — Query Endpoint
- Vector Search — Create Index, Get Index, List Indexes, Query Index
Compatibility
OAuth2 is not supported. Databricks issues OAuth2 tokens from an endpoint that is specific to your workspace address, and Secure Vault only connects to providers with a fixed sign-in address. Set Authentication to Access Token and connect a Personal Access Token instead. A workflow that uses the OAuth2 option is rejected when you publish it, with a message telling you to switch.
Execute Query waits up to about 3 minutes. A single run has a total time limit, so the node stops waiting for a long-running query and reports a timeout rather than letting the whole run expire. If your query regularly takes longer, run it against a larger warehouse, narrow the result, or write the result to a table and read it in a second agent.
All other operations are supported.