Integration · Data warehouse
Databricks + OrchestrAI
Catalog exported 2026-09-02 · Databricks website
Start, create, and terminate Databricks clusters and run jobs from chat, with cluster changes confirmed.
OrchestrAI exposes 8 Databricks operations: 3 are low-risk (read-only or low-impact), and 5 create or modify resources and run only after you confirm the plan. 2 of them also carry a step-level approval gate.
What teams use it for
Analytics platform teams use OrchestrAI to run an existing Databricks job, check its status, and stop a stuck run without switching to the workspace UI. Starting a stopped cluster and running a job are medium risk, creating a new cluster is high and confirmed, and terminating a cluster waits for approval so nobody loses an interactive session by mistake. There is no operation for notebooks, SQL warehouses, or Unity Catalog, so those stay in the Databricks workspace.
Every Databricks operation, with its risk level
| Operation | What it does | Risk | Step-level approval |
|---|---|---|---|
Get Databricks Job Run |
Get job run status | Low risk | No |
List Databricks Clusters |
List all Databricks clusters | Low risk | No |
List Databricks Jobs |
List all Databricks jobs | Low risk | No |
Cancel Databricks Run |
Cancel running Databricks job | Creates resources | No |
Run Databricks Job |
Run existing Databricks job | Creates resources | No |
Start Databricks Cluster |
Start stopped Databricks cluster | Creates resources | No |
Terminate Databricks Cluster |
Terminate Databricks cluster | Creates resources | Yes |
Create Databricks Cluster |
Create new Databricks cluster | Modifies existing | Yes |
Risk tiers come from the catalog: low is read-only or low-impact, medium creates resources and is reversible, high modifies existing resources, destructive may lose data. Every plan that creates or changes resources is shown with its cost estimate and waits for your confirmation. Operations marked with a step-level approval pause again on their own step. Destructive operations require a typed risk phrase.
What you connect
A Databricks credential (stored as databricks).
Connected-service tokens are envelope-encrypted with a per-record key wrapped by a cloud KMS.
Prompts that work
- Run the Databricks job nightly_feature_build and tell me when it finishes
- Start the shared-analytics cluster in the prod workspace
- Terminate the ml-experiments cluster that has been idle since Friday
Before anything runs
Every mutation shows its plan, cost estimate, and blast radius, then waits for your confirmation. Destructive operations require a typed risk phrase. Credentials are minted per run through OIDC federation and discarded afterward; nothing you create here is invisible later, because every resource lands in the desired-state ledger where drift is detected and can be converged. Details on the security page.
Frequently asked questions
- Can OrchestrAI create a Databricks cluster?
- Yes. Creating a cluster is a high-risk operation because it changes workspace capacity and cost, so OrchestrAI shows the node type and size and waits for confirmation.
- Will OrchestrAI cancel a Databricks job run without asking?
- Cancelling a run is medium risk and is not marked for confirmation, so it runs when you ask. Terminating a cluster, by contrast, does wait for approval.
- How does OrchestrAI authenticate to Databricks?
- You add a Databricks credential once in the connections screen. It is envelope-encrypted with a per-record key wrapped by a cloud KMS and is only decrypted inside the run that needs it.
Related integrations
Try it on your own account
Connect your cloud read-only and see your resources, drift, and costs before anything runs. $5 minimum to start. Unused credits refunded in your first 14 days.
Unused credits refunded in your first 14 days.