Actions35
- Genie Actions
- Databricks SQL Actions
- Unity Catalog Actions
- Model Serving Actions
- Files Actions
- Vector Search Actions
Overview
This node integrates with the Databricks API, specifically providing functionality to manage model serving endpoints. The "Delete Serving Endpoint" operation allows users to remove an existing model serving endpoint by specifying its name. This is useful in scenarios where a deployed model endpoint is no longer needed or must be replaced, helping maintain a clean and efficient deployment environment.
Practical examples include:
- Automatically cleaning up old or deprecated model endpoints as part of a CI/CD pipeline.
- Removing test or temporary endpoints after validation.
- Managing resource usage by deleting unused serving endpoints.
Properties
| Name | Meaning |
|---|---|
| Endpoint Name | Name of the serving endpoint to delete |
Output
The node outputs JSON data representing the response from the Databricks API after attempting to delete the specified serving endpoint. This typically includes confirmation of deletion or error details if the operation failed.
No binary data output is involved in this operation.
Dependencies
- Requires an active Databricks account with appropriate permissions to manage model serving endpoints.
- Needs an API authentication token (API key or bearer token) configured in the node credentials to authorize requests.
- The node uses the base URL provided in the credentials to connect to the Databricks workspace.
Troubleshooting
Common issues:
- Providing an incorrect or non-existent endpoint name will result in an error indicating that the endpoint could not be found.
- Insufficient permissions or invalid API tokens will cause authorization errors.
- Network connectivity problems can prevent the node from reaching the Databricks API.
Error messages and resolutions:
- "Endpoint not found": Verify the endpoint name is correct and exists.
- "Unauthorized" or "Forbidden": Check that the API token has sufficient privileges and is correctly configured.
- Timeout or network errors: Ensure stable internet connection and that the Databricks host URL is reachable.