Databricks

About Databricks

Databricks provides a workspace-level REST API for programmatic access to its data and AI platform. The API is organized around resources including compute (clusters, cluster policies, instance pools, global init scripts), jobs and pipelines, machine learning (experiments, model registry, serving endpoints), Unity Catalog (catalogs, schemas, tables, volumes, credentials, grants), Databricks SQL (warehouses, queries, alerts, dashboards), Delta Sharing, identity and access management, and workspace settings.

Requests follow REST conventions using standard HTTP methods (GET, POST, PATCH, DELETE) against versioned paths such as /api/2.0/clusters/get, scoped to a workspace instance. Databricks recommends OAuth for authentication, supporting both interactive user access and unattended service-principal access, with personal access tokens available as an alternative. Rate limits are enforced per endpoint and per workspace, with excess requests returning a 429 status.

The API is intended for developers integrating with Databricks programmatically, and is also consumed by Databricks' own developer tools, including its command-line interface, SDKs, and Terraform provider. Reference pages document request and response payloads, with examples for some endpoints covering the CLI, Terraform, and SDKs.

Key features

• REST API covering clusters, jobs, pipelines, and workspace management
• Unity Catalog APIs for catalogs, schemas, tables, and permissions
• OAuth-based authentication for both interactive users and service principals
• Databricks CLI, SDKs, and Terraform provider for common language integrations
• Vector Search and Model Serving endpoints for ML workloads
• Delta Sharing APIs for managing data providers, recipients, and shares

Frequently asked questions