Databricks
by Databricks, Inc.
Run notebooks, trigger jobs, execute SQL queries, and manage clusters on the Databricks Lakehouse platform
How agents use Databricks
- ✓Agent triggers a Databricks ETL notebook to process new data files and reports the outcome
- ✓Agent runs SQL queries against Delta Lake tables to retrieve business metrics for a report
- ✓Agent monitors job run status and retries failed jobs or sends alerts after consecutive failures
- ✓Agent starts a cluster, runs a data transformation job, and terminates the cluster to optimize costs
- ✓Agent queries aggregated analytics data from Databricks to answer natural language questions about business performance
Agent actions
Inputs: notebookPath, clusterId, parameters, timeoutSeconds
Returns: runId, state, result, runPageUrl, durationMs
Inputs: jobId, notebookParams, pythonParams, jarParams
Returns: runId, numberInJob
Inputs: runId
Returns: state, result, startTime, endTime, runPageUrl
Inputs: sql, warehouseId, catalog, schema, rowLimit
Returns: rows, columns, rowCount, statementId
Inputs: limit, nameFilter
Returns: jobs, total
Returns: clusters
Inputs: clusterId
Returns: success
Inputs: jobId, limit, activeOnly
Returns: runs, hasMore
Example workflows
Scheduled ETL pipeline
Trigger a Databricks notebook to process daily data and wait for successful completion
Ad-hoc data query
Run SQL against Delta Lake tables in Unity Catalog to answer data questions on demand
About Databricks
- Vendor
- Databricks, Inc.
- Pricing Always review details with the vendor
- Paid — Usage-based pricing on DBUs (Databricks Units). Community edition available. AWS/Azure/GCP pricing varies.
- Authentication
- Bearer token
- Rate limit Always review details with the vendor
- 200 requests / minute
- Compatible nodes
- AgentResource
- Website
- https://www.databricks.com
Build an AI workflow with Databricks
Use the Agentic Planner to design, visualize, and connect Databricks with your other tools.
Open Agentic Planner