Databricks MCP Server

Create an execution context on a running cluster. An execution context is a stateful session on a cluster where commands can be executed. Variables and state persist across commands within the same context. The cluster must be in the RUNNING state. Args: cluster_id: The unique identifier of a running cluster. language: Programming language for the context. One of "python", "sql", "scala", or "r". Defaults to "python". Returns: JSON object with the context details including id (the context_id to use in subsequent commands).

Create a new MLflow experiment. Creates an experiment as a container for ML training runs. The name must be unique and typically follows a path-like convention (e.g. "/Users/user@example.com/my-experiment"). Args: name: Unique name for the experiment. Workspace experiments typically use paths like "/Users/email/experiment-name" or "/Shared/team/experiment-name". artifact_location: Optional root artifact URI for the experiment. If not provided, the default workspace artifact location is used. Example: …

Create a new external location in the Unity Catalog metastore. Maps a cloud storage URL to a storage credential, enabling Unity Catalog to govern access to that storage path. The caller must be a metastore admin or have the CREATE_EXTERNAL_LOCATION privilege. Args: name: Name for the new external location. Must be unique. url: Cloud storage URL that this external location represents (e.g. "s3://my-bucket/path" or "abfss://container@account.dfs.core.windows.net/"). credential_name: Name of the storage credential to use for access. comment: Optional human-readable des…

Create a new Git credential for the current user. Stores authentication details for a Git provider. This credential will be used when cloning repos or syncing Git folders from the specified provider. Args: git_provider: The Git provider name. Common values: "gitHub", "gitLab", "bitbucketCloud", "bitbucketServer", "azureDevOpsServices", "awsCodeCommit". git_username: The Git username for authentication. personal_access_token: A personal access token or app password for the Git provider. This is stored …

Create a new global init script. Global init scripts run on every cluster in the workspace at startup. The script content must be base64-encoded. Scripts are disabled by default and can be positioned to control execution order. Args: name: Display name for the init script. Must be unique within the workspace. script: The script content, base64-encoded. For example, a bash script that installs a library. enabled: Whether the script should be active. Defaults to False so you can review before enabling. position: Execution order position (0-bas…

Create a new group in the workspace. The caller must be a workspace admin. Groups are used to organize users and service principals for permission management. Args: display_name: The display name for the new group. Must be unique within the workspace. Returns: JSON object with the created group's details including its assigned numeric id.

Create a new instance pool. Instance pools reduce cluster start and auto-scaling times by maintaining a set of idle, ready-to-use cloud instances. Clusters can be configured to use an instance pool for faster provisioning. Args: instance_pool_name: Display name for the instance pool. node_type_id: Instance type for pool instances (e.g. "i3.xlarge", "Standard_DS3_v2"). min_idle_instances: Minimum number of idle instances to maintain in the pool. Defaults to 0. max_capacity: Maximum total instances the pool can hold (idle + …

Create a new job with a single task. Creates a simple job with either a notebook task or a Python script task. Exactly one of notebook_path or python_file must be provided. For more complex job configurations (multi-task, schedules, etc.), use the Databricks UI or API directly. Args: name: Human-readable name for the job. Does not need to be unique. task_key: Unique key for the task within this job. Used to identify the task in multi-task jobs and in run output. Must consist of alphanumeric characters, dashes, and underscores. notebook_path: Workspace…

Create a new Unity Catalog metastore. Creates a metastore with the given name and cloud storage root. The caller must be an account admin. Each region typically has one metastore. Args: name: Name for the new metastore. Must be unique within the account. storage_root: Cloud storage root URL for the metastore's managed data (e.g. "s3://my-bucket/metastore" or "abfss://container@account.dfs.core.windows.net/metastore"). region: Optional cloud region for the metastore (e.g. "us-east-1"). Returns: JSON object with the created metastore's detai…

Create a new online table backed by a Delta table. Creates a read-optimized replica of the source Delta table for low-latency lookups. The online table automatically syncs with the source table. Args: name: Full three-level name for the online table (e.g. "my_catalog.my_schema.my_online_table"). source_table_full_name: Full three-level name of the source Delta table (e.g. "my_catalog.my_schema.my_source_table"). primary_key_columns_json: JSON array of column names that form the primary key (e.g. '["id"]' or '["us…

Create a new DLT pipeline. Creates a pipeline that runs transformation logic defined in a notebook. The pipeline manages the lifecycle of the output Delta tables automatically. For Unity Catalog pipelines, provide both catalog and target (schema). For Hive metastore pipelines, provide only target (schema name). Args: name: Human-readable name for the pipeline. target: Target schema name where output tables are created. For Unity Catalog, this is the schema within the specified catalog. For Hive metastore, this is the database name. catalog: Unity Catalog …

Create a quality monitor on a Unity Catalog table. Sets up automated data quality monitoring that generates profile and drift metrics tables. The monitor must be refreshed to compute initial metrics. Args: table_name: Full three-level name of the table to monitor in "catalog.schema.table" format. output_schema_name: Full name of the schema where profile and drift tables will be created, in "catalog.schema" format. assets_dir: Path in the workspace where monitor dashboard and configuration assets are s…

Create a new saved SQL query. Creates a reusable SQL query that can be shared, scheduled, and used as the basis for alerts. Args: name: Display name for the query. query_text: The SQL statement body. warehouse_id: The ID of the SQL warehouse to target. description: Optional human-readable description. Returns: JSON object with the created query details including its id.

Create a new Delta Sharing recipient. Creates a recipient that can be granted access to shares. The caller must have the CREATE_RECIPIENT privilege on the metastore. Args: name: Unique name for the recipient within the metastore. authentication_type: Authentication method for the recipient. Valid values: - "TOKEN" (default): Bearer token authentication. A sharing token is generated automatically and must be sent to the recipient out of band. - "DATABRICKS":…

Create a new registered model in Unity Catalog. Registered models are containers for ML model versions. They live within a catalog and schema, and each version can be served via a model serving endpoint. Args: name: Name for the new registered model. catalog_name: Name of the parent catalog. schema_name: Name of the parent schema. comment: Optional human-readable description of the model. Returns: JSON object with the created registered model's details.

Add a Git repository to the workspace. Clones a remote Git repository into the workspace under /Repos. Args: url: The URL of the Git repository to clone. Example: "https://github.com/org/repo.git". provider: Git provider — "github" (default), "gitlab", "bitbucket", "azureDevOps", or "awsCodeCommit". path: Optional workspace path for the repo. If empty, the repo is placed at /Repos/<user>/<repo_name>. Returns: JSON object with the created repo details including its id.

Create a new MLflow run within an experiment. Creates a run to track an ML training iteration. After creation, use databricks_log_metric and databricks_log_param to record metrics and parameters, and databricks_update_run to mark the run as finished. Args: experiment_id: The numeric ID of the experiment to create the run in. run_name: Optional display name for the run. If empty, MLflow assigns a random name. start_time: Optional Unix timestamp in milliseconds for the run start time. If 0, the current time is used. Returns: JSO…

Create a new schema within a catalog. The caller must be a metastore admin or have the CREATE_SCHEMA privilege on the parent catalog. Args: name: Name for the new schema, relative to the parent catalog. catalog_name: Name of the parent catalog. comment: Optional human-readable description of the schema. Returns: JSON object with the created schema's details.

Create a new Databricks-backed secret scope. The scope name must be unique within the workspace. The caller must have workspace-level MANAGE permission or be an admin. Args: scope: Name for the new secret scope. Must be unique within the workspace and consist of alphanumeric characters, dashes, underscores, and periods. Returns: Confirmation message on success.

Create a new service principal in the workspace. Service principals are non-human identities used for automation, CI/CD, and application integrations. The caller must be a workspace admin. Args: display_name: A human-readable name for the service principal. application_id: Optional UUID for the Azure AD application associated with this service principal. If not provided, one will be auto-generated by the platform. Returns: JSON object with the created service principal's details including its assigned numeric id and applicationId.

Create a new model serving endpoint. Creates an endpoint that serves a Unity Catalog model or MLflow model. The endpoint starts provisioning immediately. This tool returns as soon as the creation request is accepted -- it does NOT wait for the endpoint to become ready (which can take several minutes). Args: name: Unique name for the endpoint. Alphanumeric characters, dashes, and underscores only. model_name: Full name of the model to serve (e.g. "catalog.schema.model" for Unity Catalog models, or "model_name" for MLflow models). model_version: Version of…

Create a new Delta Sharing share. Creates an empty share that can later have tables and schemas added to it. The caller must have the CREATE_SHARE privilege on the metastore. Args: name: Unique name for the new share within the metastore. comment: Optional human-readable description of the share and the data it will contain. Returns: JSON object with the created share's details.

Create a new storage credential in the Unity Catalog metastore. Provide cloud-specific authentication parameters for the storage credential. For AWS, provide the IAM role ARN. For Azure, provide the service principal details. The caller must be a metastore admin or have the CREATE_STORAGE_CREDENTIAL privilege. Args: name: Name for the new storage credential. Must be unique. comment: Optional human-readable description. aws_iam_role_arn: For AWS — the IAM role ARN that provides access to the target S3 bucket(s). azure_service_principal_application_id: Fo…

Create a new personal access token for the current user. The token value is only returned once at creation time and cannot be retrieved later. Store it securely. Args: comment: Optional description of the token's intended use (e.g. "CI/CD pipeline" or "local development"). lifetime_seconds: Token lifetime in seconds. Use 0 or omit for no expiry (the token never expires). Returns: JSON object with the token_value (the actual secret -- store it securely!) and token_info containing token_id, comment, and expiry_time.

Create a new user in the workspace. The caller must be a workspace admin. The user_name is typically the user's email address and must be unique within the workspace. Args: user_name: The user's email address, used as their login identity. Must be unique within the workspace. display_name: Optional human-readable display name. Defaults to the user_name if not provided. Returns: JSON object with the created user's details including their assigned numeric id.

Create a new vector search endpoint. Provisions compute resources for hosting vector search indexes. The endpoint starts provisioning immediately. This tool returns without waiting for the endpoint to become ONLINE (which can take several minutes). Args: name: Unique name for the vector search endpoint. endpoint_type: Type of endpoint to create. Currently only "STANDARD" is supported. Defaults to "STANDARD".

Create a new vector search index. Creates an index on the specified endpoint. The index type determines the required parameters: For DELTA_SYNC indexes (auto-sync from a Delta table): - source_table_name is required - Optionally specify embedding_model_endpoint_name to have Databricks compute embeddings automatically from a text column For DIRECT_ACCESS indexes (manual vector management via API): - embedding_dimension is required (e.g. 1536 for OpenAI, 768 for BGE) - No source table is needed; vectors are inserted via the API Args: name: Full three-level name for the index (catalog.s…

Create a new volume within a schema. Volumes provide governed access to cloud storage locations for data files. Args: name: Name for the new volume. catalog_name: Name of the parent catalog. schema_name: Name of the parent schema. volume_type: Type of volume — "MANAGED" (default, Databricks-managed storage) or "EXTERNAL" (external storage location). comment: Optional human-readable description of the volume. Returns: JSON object with the created volume's details.

Create a new SQL warehouse. Creates a serverless SQL warehouse with the specified configuration. The warehouse starts automatically after creation. Args: name: Display name for the warehouse. cluster_size: T-shirt size for the warehouse compute (e.g. "2X-Small", "X-Small", "Small", "Medium", "Large", "X-Large", "2X-Large", "3X-Large", "4X-Large"). max_num_clusters: Maximum number of clusters for auto-scaling (1-100). auto_stop_mins: Minutes of inactivity before auto-stop (0 to disable). Returns: JSON object with the created warehouse d…

Get the metastore assigned to the current workspace. Returns the metastore that is currently assigned to the workspace where the API call is made. Returns: JSON object with the current metastore's details including metastore_id, name, storage_root, owner, and region.

Delete a file or directory from DBFS. For directories, set recursive=True to delete all contents. Attempting to delete a non-empty directory without recursive=True will result in an error. Args: path: Absolute DBFS path to delete (e.g. "dbfs:/mnt/data/old_file.csv"). recursive: If True, recursively delete all files and subdirectories. Returns: Confirmation message on success.

Download a file from DBFS and return its content as base64. Reads the file at the given DBFS path and returns the content encoded as a base64 string. Suitable for files up to a few MB. Args: path: Absolute DBFS path of the file to download (e.g. "dbfs:/mnt/data/report.csv"). Returns: JSON object with 'path', 'content_base64' (the base64-encoded file content), and 'size_bytes'.

Get the status and metadata of a single DBFS path. Returns information about whether the path is a file or directory, its size, and modification time. Args: path: Absolute DBFS path to check (e.g. "dbfs:/mnt/data/file.csv"). Returns: JSON object with path, is_dir, file_size, and modification_time.

List files and directories at a DBFS path. Returns metadata for each entry including file name, path, size, and whether it is a directory. Args: path: Absolute DBFS path to list (e.g. "dbfs:/mnt/data" or "/mnt/data"). Returns: JSON array of file status objects with path, is_dir, file_size, and modification_time. Results are capped at 100 items.

Create a directory in DBFS, including any necessary parent directories. Works like "mkdir -p" -- creates the full directory tree if intermediate directories do not exist. No error if the directory already exists. Args: path: Absolute DBFS path of the directory to create (e.g. "dbfs:/mnt/data/output"). Returns: Confirmation message on success.

Upload a file to DBFS from base64-encoded content. The content must be provided as a base64-encoded string. This is suitable for files up to a few MB. For very large files, consider using the DBFS streaming API directly. Args: path: Absolute DBFS destination path (e.g. "dbfs:/mnt/data/upload.csv"). contents_base64: The file content encoded as a base64 string. overwrite: If True, overwrite any existing file at the path. Returns: Confirmation message on success.

Delete a Databricks App. Permanently deletes the app and all its deployments. This action cannot be undone. The app's URL will become unavailable immediately. Args: name: The name of the app to delete. Returns: Confirmation that the app has been deleted, or an error if the app was not found.

Delete a catalog from the Unity Catalog metastore. The caller must be the catalog owner or a metastore admin. Args: name: Name of the catalog to delete. force: If True, delete the catalog even if it contains schemas. All child schemas (and their tables) will also be deleted. Returns: Confirmation message on success.

Delete an external connection. Permanently removes the connection and its stored credentials. Foreign catalogs that depend on this connection will no longer function. The caller must be the connection owner or a metastore admin. Args: name: Name of the connection to delete. Returns: Confirmation message on success.

Delete a Lakebase PostgreSQL database instance. Deletes the specified database instance. By default this is a hard delete (purge=True). Use force=True to also delete any descendant instances. WARNING: This permanently destroys the database and all its data. This action cannot be undone when purge is True. Args: name: The name of the database instance to delete. force: If True, also delete any descendant instances. Defaults to False, which prevents deletion if descendants exist. purge: If True (default), hard delete the instance permanently. If False, soft d…

Mark an MLflow experiment as deleted (soft delete). The experiment is moved to a trash state and can be restored within the retention period. All runs within the experiment are also marked as deleted. The caller must be the experiment owner or a workspace admin. Args: experiment_id: The numeric ID of the experiment to delete. Returns: Confirmation message on success.

Delete an external location from the Unity Catalog metastore. The caller must be the external location owner or a metastore admin. Args: name: Name of the external location to delete. force: If True, delete even if the location is used by managed tables or volumes. Returns: Confirmation message on success.

Delete a user-defined function from Unity Catalog. The caller must be the function owner or a metastore admin. Args: full_name: Full three-level name of the function in "catalog.schema.function" format (e.g. "my_catalog.my_schema.my_function"). Returns: Confirmation message on success.

Delete a Git credential. Removes the stored authentication details. Any Repos or Git folders using this credential will no longer be able to sync until a new credential is configured. Args: credential_id: The unique numeric identifier of the credential to delete. Returns: Confirmation message on success.

Delete a global init script from the workspace. The script will no longer run on new cluster starts. Clusters that are already running are not affected until they are restarted. Args: script_id: The unique identifier of the script to delete. Returns: Confirmation message on success.

Delete a group from the workspace. The caller must be a workspace admin. This removes the group but does not delete its member users or service principals. Args: group_id: The numeric ID of the group to delete. Returns: Confirmation message on success.

Delete an instance pool. Permanently removes the instance pool and terminates all idle instances. Clusters currently using this pool are not affected but will not be able to scale using the pool after deletion. Args: instance_pool_id: The unique identifier of the instance pool to delete. Returns: Confirmation message on success.

Delete a job and all its associated runs. Permanently removes the job definition. Active runs are cancelled. This action cannot be undone. Args: job_id: The unique numeric identifier of the job to delete.

Delete a Unity Catalog metastore. This is a destructive operation. If force=True, the metastore is deleted even if it contains catalogs, schemas, or tables. The caller must be an account admin. Args: id: The unique identifier (UUID) of the metastore to delete. force: If True, forcibly delete the metastore and all its contents. Returns: Confirmation message on success.

Delete an online table. Removes the online table and stops syncing from the source Delta table. The source Delta table is not affected. Args: name: Full three-level name of the online table to delete (e.g. "my_catalog.my_schema.my_online_table"). Returns: Confirmation message on success.

Delete a DLT pipeline. Permanently removes the pipeline definition. If the pipeline is running, it is stopped first. The output Delta tables created by the pipeline are NOT deleted. This action cannot be undone. Args: pipeline_id: The unique identifier of the pipeline to delete.

Delete the quality monitor from a table. Removes the monitor configuration. The profile and drift metric tables in the output schema are NOT deleted and can be cleaned up separately if desired. Args: table_name: Full three-level name of the monitored table in "catalog.schema.table" format. Returns: Confirmation message on success.

Delete a Delta Sharing recipient. Permanently removes the recipient. They will no longer be able to access any shares they were previously granted. This action cannot be undone. Args: name: The name of the recipient to delete. Returns: Confirmation message on success.

Delete a registered model from Unity Catalog. Permanently removes the registered model and all its versions. The caller must be the model owner or a metastore admin. This action cannot be undone. Args: full_name: Full three-level name of the registered model in "catalog.schema.model" format (e.g. "my_catalog.my_schema.my_model"). Returns: Confirmation message on success.

Mark an MLflow run as deleted (soft delete). The run is moved to a trash state and can be restored within the retention period. The caller must be the run owner or a workspace admin. Args: run_id: The UUID of the run to delete. Returns: Confirmation message on success.

Delete a schema from Unity Catalog. The caller must be the schema owner or an owner of the parent catalog. Args: full_name: Full name of the schema in "catalog.schema" format (e.g. "my_catalog.my_schema"). Returns: Confirmation message on success.

Delete a secret from a scope. The caller must have WRITE or MANAGE permission on the scope. Args: scope: Name of the secret scope containing the secret. key: Name of the secret key to delete. Returns: Confirmation message on success.

Delete a secret scope and all of its secrets. This is a destructive operation -- all secrets within the scope will be permanently deleted. The caller must have MANAGE permission on the scope. Args: scope: Name of the secret scope to delete. Returns: Confirmation message on success.

Delete a service principal from the workspace. The caller must be a workspace admin. This removes the service principal's access but does not delete resources it owns. Args: id: The numeric ID of the service principal to delete. Returns: Confirmation message on success.

Delete a serving endpoint. Permanently removes the endpoint and releases all associated compute resources. This action cannot be undone. Args: name: Name of the serving endpoint to delete.

Delete a Delta Sharing share. Permanently removes the share. Recipients who had access to this share will no longer be able to query its tables. The underlying tables are not affected. Args: name: The name of the share to delete. Returns: Confirmation message on success.

Delete a storage credential from the Unity Catalog metastore. The credential must not be referenced by any external locations unless force=True is specified. The caller must be the credential owner or a metastore admin. Args: name: Name of the storage credential to delete. force: If True, delete even if referenced by external locations. Returns: Confirmation message on success.

Delete a table from Unity Catalog. The caller must be the table owner or a metastore admin. Args: full_name: Full three-level name of the table in "catalog.schema.table" format (e.g. "my_catalog.my_schema.my_table"). Returns: Confirmation message on success.

Revoke (delete) a personal access token for the current user. The token will immediately stop working for authentication. Args: token_id: The unique identifier of the token to revoke. Returns: Confirmation message on success.

Revoke any user's personal access token (admin only). Workspace admins can use this to revoke tokens belonging to any user. The token will immediately stop working for authentication. Args: token_id: The unique identifier of the token to revoke. Returns: Confirmation message on success.

Delete a user from the workspace. The caller must be a workspace admin. This removes the user's access but does not delete resources they own. Args: user_id: The numeric ID of the user to delete. Returns: Confirmation message on success.

Delete a vector search endpoint. Permanently removes the endpoint and all its compute resources. All indexes hosted on this endpoint must be deleted first, or the operation will fail. Args: name: Name of the vector search endpoint to delete.

Delete a vector search index. Permanently removes the index and its data. For DELTA_SYNC indexes, the source Delta table is NOT affected. This action cannot be undone. Args: name: Full three-level name of the index to delete (catalog.schema.index_name).

Delete a volume from Unity Catalog. The caller must be the volume owner or a metastore admin. Args: full_name: Full three-level name of the volume in "catalog.schema.volume" format (e.g. "my_catalog.my_schema.my_volume"). Returns: Confirmation message on success.

Delete a SQL warehouse. Permanently deletes the warehouse. This action cannot be undone. Args: id: The unique identifier of the SQL warehouse to delete. Returns: Confirmation message on success.

Delete a workspace object or directory. Deletes a notebook, file, or directory at the given path. For directories, use recursive=True to delete all contents. Args: path: Absolute workspace path of the object to delete. recursive: If True, recursively delete all contents of a directory. Required to be True for non-empty directories. Returns: Confirmation message on success.

Deploy source code to a Databricks App. Creates a new deployment for the specified app. The source code at the given workspace path will be packaged and deployed. This is an asynchronous operation. The deployment will proceed in the background. Use databricks_get_app or databricks_list_app_deployments to monitor progress. Args: name: The name of the app to deploy to. source_code_path: The workspace filesystem path of the source code (e.g., "/Workspace/Users/me@company.com/my-app"). mode: Deployment mode. "SNAPSHOT" (default) creates an immutable copy …

Destroy an execution context on a cluster. Releases the resources associated with the context. Any running commands in the context will be cancelled. It is good practice to destroy contexts when they are no longer needed. Args: cluster_id: The unique identifier of the cluster. context_id: The execution context ID to destroy. Returns: Confirmation message on success.

Edit the configuration of an existing compute cluster. Updates the cluster definition. The cluster may be restarted to apply changes if it is currently running. This call returns immediately without blocking. Args: cluster_id: The unique identifier of the cluster to edit. cluster_name: New display name for the cluster. spark_version: Spark runtime version string (e.g. "14.3.x-scala2.12", "15.0.x-scala2.12"). node_type_id: Instance type for worker nodes (e.g. "i3.xlarge", "Standard_DS3_v2"). num_workers: Number of worker nodes. Use 0 fo…

Execute a command in an existing execution context. Submits a command for asynchronous execution on the cluster. Returns immediately with a command ID that can be polled for status and results using databricks_command_status. Args: cluster_id: The unique identifier of the cluster. context_id: The execution context ID returned by databricks_create_execution_context. language: Programming language of the command. One of "python", "sql", "scala", or "r". command: The code to execute (e.g. "print('hello')" for Python or "SELECT 1" for SQ…

Execute a SQL statement on a Databricks SQL warehouse. Runs the statement synchronously with a 50-second wait timeout and returns inline JSON results. For long-running queries, use databricks_get_statement_status to poll for completion. Args: warehouse_id: The ID of the SQL warehouse to execute on. statement: The SQL statement to execute (max 16 MiB). catalog: Optional default catalog for the statement (like "USE CATALOG"). schema: Optional default schema for the statement (like "USE SCHEMA"). Returns: JSON object with the statement result includ…

Export a notebook from the workspace. Exports the notebook source content. The content is returned as a base64-encoded string. Args: path: Absolute workspace path of the notebook to export. Example: "/Users/user@example.com/my_notebook". Returns: JSON object with the exported content (base64-encoded) and the notebook path.

Export the content of a job run. Exports the run's notebook content (for notebook tasks) in HTML format. This includes cell outputs, visualizations, and any dashboard views generated during the run. Args: run_id: The unique numeric identifier of the run to export.

Create a directory in a Unity Catalog Volume. Creates the directory at the specified path. Parent directories must already exist. Args: directory_path: Volume-relative directory path to create (e.g. "/Volumes/my_catalog/my_schema/my_volume/output"). Returns: Confirmation message on success.

Delete a file from a Unity Catalog Volume. Permanently removes the file at the specified path. This operation cannot be undone. Args: file_path: Volume-relative file path to delete (e.g. "/Volumes/my_catalog/my_schema/my_volume/data/old.csv"). Returns: Confirmation message on success.

Download a file from a Unity Catalog Volume as base64-encoded content. Reads the file at the given Volume path and returns the content encoded as a base64 string. Suitable for files up to a few MB. Args: file_path: Volume-relative file path to download (e.g. "/Volumes/my_catalog/my_schema/my_volume/data/report.csv"). Returns: JSON object with 'path', 'content_base64' (the base64-encoded file content), and 'size_bytes'.

Get metadata for a file in a Unity Catalog Volume. Returns metadata including content type and content length for the specified file. Does not return the file contents. Args: file_path: Volume-relative file path (e.g. "/Volumes/my_catalog/my_schema/my_volume/data/file.csv"). Returns: JSON object with file metadata including content_type and content_length.

List files and directories within a Unity Catalog Volume path. Returns the contents of a directory under a UC Volume. The path should be in the Volumes path format. Args: directory_path: Volume-relative directory path (e.g. "/Volumes/my_catalog/my_schema/my_volume/data"). Returns: JSON array of directory entry objects with path, name, is_directory, and file_size. Results are capped at 100 items.

Upload a file to a Unity Catalog Volume from base64-encoded content. The content must be provided as a base64-encoded string. The file will be created at the specified path, overwriting any existing file. Args: file_path: Volume-relative destination path (e.g. "/Volumes/my_catalog/my_schema/my_volume/data/upload.csv"). contents_base64: The file content encoded as a base64 string. Returns: Confirmation message on success including bytes written.

Find an available SQL warehouse and start it if needed. Looks for running warehouses first. If none are running, finds a stopped warehouse and starts it. Returns the warehouse ID ready for queries.

Generate temporary PostgreSQL credentials for Lakebase instances. Generates short-lived credentials that can be used to connect directly to the PostgreSQL instances using standard PG clients (psql, psycopg, etc.). The returned credentials include the username, password, and connection parameters needed for direct database access. Credentials are temporary and will expire after a set period. Args: instance_names: Comma-separated list of database instance names to scope the credential to (e.g., "my-instance" or "instance-1,instance-2"). Returns: …

Send a follow-up message in an existing Genie conversation. Adds a new question or follow-up to an existing Genie conversation. Genie uses all previous messages in the conversation as context when generating its response, enabling multi-turn data exploration. This is an asynchronous operation. Use databricks_genie_get_message to poll for the answer. Args: space_id: The ID of the Genie Space containing the conversation. conversation_id: The ID of the existing conversation to continue. content: The follow-up question or refinement (e.g., "Break that down by region" or…

Re-execute the SQL query from a Genie message. Re-runs the SQL query that was generated for a specific Genie message. This is useful when you want to get fresh results for a previously asked question, or when the original query execution timed out. Args: space_id: The ID of the Genie Space containing the conversation. conversation_id: The ID of the conversation containing the message. message_id: The ID of the message whose query to re-execute. Returns: JSON object with the re-executed query result including updated column schema and row data.

Get a Genie message and its response. Retrieves a specific message from a Genie conversation, including the AI-generated response, any SQL query that was generated, and the status of the message processing. Poll this endpoint after starting a conversation or sending a message to check if the response is ready. The message status will indicate whether processing is complete. Args: space_id: The ID of the Genie Space containing the conversation. conversation_id: The ID of the conversation containing the message. message_id: The ID of the specific message to retrieve. Returns: …

Get the SQL query result from a Genie message. Retrieves the tabular data result from a Genie message's generated SQL query. Use this after a message has been processed to get the actual data rows returned by the query. Note: This requires the message to have completed processing and to have generated a SQL query with results. Check the message status first with databricks_genie_get_message. Args: space_id: The ID of the Genie Space containing the conversation. conversation_id: The ID of the conversation containing the message. message_id: The ID of the message whose query resul…

Start a new Genie AI/BI conversation with a question. Initiates a new conversation in the specified Genie Space by asking a natural language question about your data. Genie will interpret the question, generate SQL, and produce an answer. This is an asynchronous operation. The response includes conversation and message IDs that can be used to retrieve the result once processing completes. Use databricks_genie_get_message to poll for the answer. Args: space_id: The ID of the Genie Space to start the conversation in. Genie Spaces are configured with specific data sources …

Get detailed information about a specific Databricks App. Retrieves the full configuration and status of an app including its active deployment, compute status, URL, creator, and resources. Args: name: The name of the app. Must be a lowercase alphanumeric string with hyphens (e.g., "my-dashboard-app"). Returns: JSON object with complete app details including name, description, url, compute_status, active_deployment, and pending_deployment.

Get details of a specific app deployment. Retrieves complete information about a single deployment including its status, source code path, deployment mode, and any error messages. Args: app_name: The name of the app. deployment_id: The unique identifier of the deployment to retrieve. Returns: JSON object with full deployment details including status, source_code_path, mode, create_time, and update_time.

Get the environment configuration for a Databricks App. Retrieves the app's environment variables, secrets, and resource bindings. Useful for understanding what configuration an app has access to at runtime. Args: name: The name of the app whose environment to retrieve. Returns: JSON object with environment variables, secret references, and resource bindings configured for the app.

Get detailed information about a specific catalog. Args: name: The name of the catalog to retrieve (e.g. "my_catalog"). Returns: JSON object with catalog details including name, owner, comment, created_at, updated_at, and properties.

Get detailed information about a specific compute cluster. Args: cluster_id: The unique identifier of the cluster. Returns: JSON object with full cluster details including cluster_name, state, state_message, spark_version, node_type_id, driver_node_type_id, num_workers, autoscale, spark_conf, and runtime information.

Get detailed information about a specific connection. Args: name: The name of the connection to retrieve. Returns: JSON object with full connection details including name, connection_type, options (host, port, etc.), owner, comment, and timestamps.

Get information about the currently authenticated user. Returns the identity of the user or service principal making the API call. Useful for verifying authentication and determining the caller's user ID, name, and group memberships. Returns: JSON object with the current user's details including id, userName, displayName, active status, and groups.

Get detailed information about a specific Lakeview dashboard. Retrieves the full dashboard configuration including its display name, serialized layout, datasets, warehouse binding, and lifecycle state. Args: dashboard_id: The UUID identifying the dashboard (e.g., "01ef1234-5678-abcd-ef90-1234567890ab"). Returns: JSON object with complete dashboard details including dashboard_id, display_name, serialized_dashboard (layout JSON), warehouse_id, path, and lifecycle_state.

Get detailed information about a specific Lakebase database instance. Retrieves the full configuration, status, and connection details for a database instance. Includes hostname, port, and connection strings needed to connect to the PostgreSQL instance. Args: name: The name of the database instance to retrieve. Returns: JSON object with instance name, state, hostname, port, connection_string, jdbc_connection_url, and capacity details.

Get the effective (inherited + direct) permissions on a securable. Returns the effective permissions for the calling user on the specified securable, including permissions inherited from parent objects. Args: securable_type: The type of the securable object. Must be one of: "CATALOG", "SCHEMA", "TABLE", "VOLUME", "FUNCTION", "REGISTERED_MODEL", "EXTERNAL_LOCATION", "STORAGE_CREDENTIAL", "METASTORE", "CONNECTION", "SHARE", "RECIPIENT", "PROVIDER". full_name: The full name of the securable. Format depends o…

Get detailed information about a specific MLflow experiment. Args: experiment_id: The numeric ID of the experiment to retrieve (e.g. "123456789"). Returns: JSON object with full experiment details including experiment_id, name, artifact_location, lifecycle_stage, tags, and timestamps.

Get detailed information about a specific MLflow run. Args: run_id: The UUID of the run to retrieve (e.g. "a1b2c3d4e5f6a7b8c9d0e1f2a3b4c5d6"). Returns: JSON object with full run details including run_id, experiment_id, status, start_time, end_time, metrics, params, tags, and artifact_uri.

Get detailed information about a specific external location. Args: name: The name of the external location to retrieve. Returns: JSON object with external location details including name, url, credential_name, owner, created_at, and updated_at.

Get detailed information about a specific function. Args: full_name: Full three-level name of the function in "catalog.schema.function" format (e.g. "my_catalog.my_schema.my_function"). Returns: JSON object with function details including name, input_params, data_type, full_data_type, and routine_body.

Get detailed information about a specific Git credential. Args: credential_id: The unique numeric identifier of the Git credential. Returns: JSON object with credential details including credential_id, git_provider, and git_username.

Get detailed information about a specific global init script. Returns the full script metadata including its base64-encoded content. Args: script_id: The unique identifier of the global init script. Returns: JSON object with script details including script_id, name, enabled, position, script (base64-encoded content), created_at, created_by, updated_at, and updated_by.

Get the permissions granted on a Unity Catalog securable. Returns all privilege assignments for the specified securable object, showing which principals have which privileges. Args: securable_type: The type of the securable object. Must be one of: "CATALOG", "SCHEMA", "TABLE", "VOLUME", "FUNCTION", "REGISTERED_MODEL", "EXTERNAL_LOCATION", "STORAGE_CREDENTIAL", "METASTORE", "CONNECTION", "SHARE", "RECIPIENT", "PROVIDER". full_name: The full name of the securable. Format depends on type: - CA…

Get detailed information about a specific group. Args: group_id: The numeric ID of the group to retrieve. Returns: JSON object with full group details including id, displayName, members, roles, and entitlements.

Get detailed information about a specific instance pool. Args: instance_pool_id: The unique identifier of the instance pool. Returns: JSON object with pool details including instance_pool_name, node_type_id, min_idle_instances, max_capacity, idle_instance_autotermination_minutes, and stats about pending/used/idle instances.

Get detailed information about a specific job. Returns the full job definition including all tasks, cluster configuration, schedule, email notifications, and access control settings. Args: job_id: The unique numeric identifier of the job.

Get detailed information about a specific metastore. Args: id: The unique identifier (UUID) of the metastore. Returns: JSON object with metastore details including metastore_id, name, storage_root, default_data_access_config_id, owner, region, created_at, and updated_at.

Get detailed information about a specific online table. Returns the online table's configuration, sync status, and metadata. Args: name: Full three-level name of the online table (e.g. "my_catalog.my_schema.my_online_table"). Returns: JSON object with online table details including name, spec (source table, primary key columns, sync mode), status, and table_serving_url.

Get the permission levels available for a workspace object. Returns all possible permission levels that can be granted on the specified object type. Use this to discover valid permission levels before setting permissions. Args: object_type: The type of object to query permission levels for. Common values: "clusters", "cluster-policies", "directories", "experiments", "jobs", "notebooks", "pipelines", "registered-models", "repos", "serving-endpoints", "sql/warehouses", "tokens". object_id: The ID of th…

Get the access control list (ACL) for a workspace object. Returns the current permissions assigned to users, groups, and service principals for the specified object. Args: object_type: The type of object to query permissions for. Common values: "clusters", "cluster-policies", "directories", "experiments", "jobs", "notebooks", "pipelines", "registered-models", "repos", "serving-endpoints", "sql/warehouses". object_id: The ID of the specific object. Format varies by object type. Returns: JSON object…

Get detailed information about a specific DLT pipeline. Returns the full pipeline configuration including name, target schema, catalog, libraries, clusters, current state, last update status, and health metrics. Args: pipeline_id: The unique identifier of the pipeline (UUID format).

Get detailed information about a specific Delta Sharing provider. Returns the provider definition including authentication type, recipient profile, and available shares. Args: name: The name of the provider to retrieve. Returns: JSON object with full provider details including name, comment, authentication_type, and recipient_profile.

Get the published version of a Lakeview dashboard. Retrieves the published (live) version of a dashboard, which may differ from the current draft. Returns the published dashboard's configuration, embed credentials setting, and warehouse binding. This method may not be available in all SDK versions; if unavailable, an informative error is returned. Args: dashboard_id: The UUID identifying the dashboard whose published version to retrieve. Returns: JSON object with the published dashboard details including embed_credentials, warehouse_id, and publish timestamp.

Get the quality monitor configuration for a table. Args: table_name: Full three-level name of the monitored table in "catalog.schema.table" format. Returns: JSON object with monitor configuration including output_schema_name, assets_dir, monitor_type, schedule, and status information.

Get the status of a specific quality monitor refresh. Args: table_name: Full three-level name of the monitored table in "catalog.schema.table" format. refresh_id: The unique identifier of the refresh run. Returns: JSON object with refresh details including refresh_id, state (PENDING, RUNNING, SUCCESS, FAILED, CANCELED), start_time, and end_time.

Get detailed information about a specific Delta Sharing recipient. Returns the recipient definition including authentication type, activation status, sharing tokens, and associated shares. Args: name: The name of the recipient to retrieve. Returns: JSON object with full recipient details including name, authentication_type, comment, activation_url, and tokens.

Get detailed information about a specific registered model. Args: full_name: Full three-level name of the registered model in "catalog.schema.model" format (e.g. "my_catalog.my_schema.my_model"). Returns: JSON object with model details including name, catalog_name, schema_name, owner, comment, and version information.

Get detailed information about a specific Git repository. Args: repo_id: The numeric ID of the repository. Returns: JSON object with repo details including id, url, provider, path, branch, and head_commit_id.

Get detailed information about a specific job run. Returns the full run details including state (PENDING, RUNNING, TERMINATED, SKIPPED, INTERNAL_ERROR), task results, cluster info, start/end times, and any error messages. Args: run_id: The unique numeric identifier of the run.

Get the output of a completed job run. Returns the output produced by the run, which varies by task type: - Notebook tasks: notebook output and exit value - Python tasks: stdout/stderr logs - SQL tasks: query results Only available for completed runs. Args: run_id: The unique numeric identifier of the run.

Get detailed information about a specific schema. Args: full_name: Full name of the schema in "catalog.schema" format (e.g. "my_catalog.my_schema"). Returns: JSON object with schema details including name, catalog_name, owner, comment, and properties.

Get detailed information about a specific service principal. Args: id: The numeric ID of the service principal to retrieve. Returns: JSON object with full service principal details including id, displayName, applicationId, active status, and entitlements.

Get detailed information about a specific serving endpoint. Returns the full endpoint configuration including served entities, traffic routing, state, pending config changes, and tags. Args: name: The name of the serving endpoint. Must be unique within the workspace.

Retrieve build logs for a served model within an endpoint. Returns the build logs generated during model container creation. Useful for debugging deployment failures or checking build progress. Args: name: Name of the serving endpoint. served_model_name: Name of the served model within the endpoint whose build logs you want to retrieve.

Get the OpenAPI specification for a serving endpoint. Retrieves the auto-generated OpenAPI spec that describes the REST API for querying the endpoint. Useful for understanding the expected input/output schema of the served model. This method may not be available in all SDK versions; if unavailable, an informative error is returned. Args: name: Name of the serving endpoint.

Get detailed information about a specific share. Returns the share definition including all tables and schemas that are part of the share. Args: name: The name of the share to retrieve. Returns: JSON object with full share details including name, owner, comment, objects (shared tables and schemas), created_at, and updated_at.

Get the status and results of a previously executed SQL statement. Use this to poll for completion of long-running statements or to retrieve results after an asynchronous execution. Args: statement_id: The statement ID returned by databricks_execute_sql. Returns: JSON object with statement status, manifest, and result data if the statement has completed.

Get detailed information about a specific storage credential. Args: name: The name of the storage credential to retrieve. Returns: JSON object with credential details including name, owner, cloud-specific configuration, created_at, and updated_at.

Get detailed information about a specific table. Args: full_name: Full three-level name of the table in "catalog.schema.table" format (e.g. "my_catalog.my_schema.my_table"). Returns: JSON object with table details including columns, table_type, data_source_format, storage_location, and properties.

Get detailed information about a specific user. Args: user_id: The numeric ID of the user to retrieve (as returned by list_users or create_user). Returns: JSON object with full user details including id, userName, displayName, active status, emails, groups, and roles.

Get detailed information about a vector search endpoint. Returns the endpoint configuration, current state (PROVISIONING, ONLINE, OFFLINE), and associated metadata. Args: name: Name of the vector search endpoint.

Get detailed information about a specific vector search index. Returns the index configuration including type (DELTA_SYNC or DIRECT_ACCESS), status, endpoint, primary key, embedding source, and sync pipeline details. Args: index_name: Full three-level name of the index (catalog.schema.index_name). Returns: JSON object with full index details including name, endpoint_name, index_type, primary_key, and status.

Get detailed information about a specific SQL warehouse. Args: id: The unique identifier of the SQL warehouse. Returns: JSON object with warehouse details including name, state, cluster_size, auto_stop_mins, num_clusters, and health info.

Get the status and metadata of a workspace object. Returns detailed information about a notebook, file, directory, or other workspace object at the given path. Args: path: Absolute workspace path of the object. Example: "/Users/user@example.com/my_notebook". Returns: JSON object with path, object_type, object_id, language, created_at, modified_at, and size information.

Import a notebook or file into the workspace. Imports content as a notebook at the specified path. The content should be base64-encoded. Maximum content size is 10 MB. Args: path: Absolute workspace path where the notebook will be created. Example: "/Users/user@example.com/my_notebook". content: Base64-encoded content of the notebook. format: Import format — "SOURCE" (default), "HTML", "JUPYTER", "DBC", "R_MARKDOWN", or "AUTO". language: Notebook language — "PYTHON" (default), "SQL", "SCALA", or "R". Required when format is "SOURCE". ov…

List all SQL alerts accessible to the current user. Alerts trigger notifications when a query result meets a specified condition. Returns: JSON array of alert objects with id, name, query information, and condition details. Results are capped at 100 items.

List all deployments for a Databricks App. Returns the deployment history for the specified app, including both active and past deployments with their status, timestamps, and source code paths. Args: name: The name of the app whose deployments to list. Returns: JSON array of deployment objects with deployment_id, status, source_code_path, mode, and creation timestamps.

List all Databricks Apps in the workspace. Returns a paginated list of apps with their names, descriptions, status, and URLs. Use this to discover what apps exist in the workspace before performing operations on them. Returns: JSON array of app objects, each containing name, description, active deployment info, and compute status. Limited to 100 results.

List artifacts associated with an MLflow run. Returns the files and directories stored as artifacts for the specified run. Artifacts include model files, data files, plots, and any other files logged during the run. Args: run_id: The UUID of the run whose artifacts to list. path: Optional relative path within the artifact directory. If empty, lists artifacts at the root level. Returns: JSON array of artifact objects, each containing file_path, is_dir, and file_size.

List all catalogs in the Unity Catalog metastore. Returns a JSON array of catalog objects, each containing name, owner, comment, and other metadata. Results are capped at 100 items.

List recent events for a compute cluster. Returns events such as creation, configuration changes, state transitions, and errors. Useful for auditing and troubleshooting cluster behavior. Args: cluster_id: The unique identifier of the cluster. limit: Maximum number of events to return. Defaults to 50. Returns: JSON object with an array of cluster event objects, each containing timestamp, type, and details.

List all cluster policies in the workspace. Cluster policies constrain the attributes available during cluster creation, enforcing organizational standards and cost controls. Returns: JSON array of policy objects with policy_id, name, description, definition, and creator information. Results are capped at 100 items.

List all compute clusters in the workspace. Returns both running and terminated clusters, including all-purpose and job clusters. Each entry includes the cluster's current state. Returns: JSON array of cluster objects with cluster_id, cluster_name, state, spark_version, node_type_id, and autoscale config. Results are capped at 100 items.

List all external connections in the workspace. Returns connections configured for Lakehouse Federation and other external data access patterns. Each connection stores the type, host, port, and credential information needed to reach an external system. Returns: JSON array of connection objects, each containing name, connection_type, host, port, comment, owner, and created/updated timestamps. Results are capped at 100 items.

List all Lakeview dashboards in the workspace. Returns a paginated list of dashboards with their IDs, display names, paths, and lifecycle status. Only active (non-trashed) dashboards are returned by default. Returns: JSON array of dashboard objects with dashboard_id, display_name, path, lifecycle_state, and warehouse_id. Limited to 100 results.

List all catalogs in a Lakebase database instance. Returns the catalogs (logical databases) within the specified database instance. Each catalog can contain multiple schemas and tables. Args: instance_name: The name of the database instance whose catalogs to list. Returns: JSON array of catalog objects with names and properties.

List all Lakebase PostgreSQL database instances in the workspace. Returns all database instances with their names, states, connection info, and capacity. Use this to discover available database instances before performing operations on them. Returns: JSON array of database instance objects with instance name, state, connection details, and capacity information.

List all roles for a Lakebase database instance. Returns the database roles configured for the specified instance. Roles control access permissions within the PostgreSQL database. Args: instance_name: The name of the database instance whose roles to list. Returns: JSON array of role objects with names and permissions.

List tables in a Lakebase database schema. Returns all tables within the specified catalog and schema of a database instance. Defaults to the "public" schema. Args: instance_name: The name of the database instance. catalog_name: The name of the catalog containing the tables. schema_name: The schema to list tables from. Defaults to "public". Returns: JSON array of table objects with names, columns, and metadata.

List MLflow experiments in the workspace. Returns experiments ordered by creation time (newest first). Each experiment contains one or more runs that track ML training iterations. Args: max_results: Maximum number of experiments to return (default 25). Returns: JSON array of experiment objects, each containing experiment_id, name, artifact_location, lifecycle_stage, last_update_time, and creation_time.

List all external locations in the Unity Catalog metastore. External locations map cloud storage paths to Unity Catalog for governance. Each external location references a storage credential and a cloud storage URL. Returns: JSON array of external location objects with name, url, credential_name, owner, and metadata. Results are capped at 100 items.

List all user-defined functions within a schema. Args: catalog_name: Name of the parent catalog. schema_name: Name of the schema to list functions from. Returns: JSON array of function objects with name, catalog_name, schema_name, input_params, and return type info. Results are capped at 100 items.

List all Git credentials for the current user. Returns metadata about each stored Git credential including the provider, username, and credential ID. Returns: JSON array of Git credential objects with credential_id, git_provider, and git_username. Results are capped at 100 items.

List all global init scripts in the workspace. Global init scripts run on every cluster in the workspace at startup, in the order specified by their position. Each entry includes the script's name, enabled status, and position. Returns: JSON array of global init script objects with script_id, name, enabled, position, created_at, and updated_at. Results are capped at 100 items.

List workspace groups with optional SCIM filtering. Groups are used to manage permissions for collections of users and service principals. Args: filter_str: Optional SCIM filter expression to narrow results. Examples: - 'displayName eq "admins"' (exact match) - 'displayName co "eng"' (contains) Leave empty to list all groups. count: Maximum number of groups to return (default 100). Returns: JSON array of group objects, each containing id, displayName, members, and roles.

List all instance pools in the workspace. Instance pools reduce cluster start and auto-scaling times by maintaining a set of idle, ready-to-use cloud instances. Returns: JSON array of instance pool objects with instance_pool_id, instance_pool_name, node_type_id, min_idle_instances, and state. Results are capped at 100 items.

List jobs in the workspace. Returns job definitions including name, schedule, tasks, and cluster configuration. Results are paginated; use the limit parameter to control how many jobs are returned. Args: limit: Maximum number of jobs to return. Must be between 1 and 100. Defaults to 25. name: Optional filter to return only jobs whose name contains this string (case-insensitive partial match).

List all Unity Catalog metastores accessible to the caller. Returns metastore metadata including name, storage root, owner, and region. Typically there is one metastore per region. Returns: JSON array of metastore objects with metastore_id, name, storage_root, owner, region, and other metadata. Results are capped at 100 items.

List all versions of a Unity Catalog registered model. Returns version numbers, creation timestamps, current stage, and source run information for each version of the specified model. Useful for finding available versions before creating a serving endpoint. Args: full_model_name: Full three-level name of the Unity Catalog model in "catalog.schema.model" format.

List all available node types for compute clusters. Returns the available instance types that can be used as worker or driver nodes when creating or editing clusters. Each entry includes instance type ID, memory, CPU cores, and availability. Returns: JSON object with available node types and their specifications.

List recent events for a DLT pipeline. Returns pipeline events in reverse chronological order. Events include update starts/completions, data quality violations, errors, schema changes, and flow progress. Useful for monitoring and debugging pipeline behavior. Args: pipeline_id: The unique identifier of the pipeline. max_results: Maximum number of events to return. Defaults to 25.

List DLT pipelines in the workspace. Returns summary information for each pipeline including name, ID, state (IDLE, RUNNING, FAILED), target schema, and creation time. Args: max_results: Maximum number of pipelines to return. Defaults to 25.

List all Delta Sharing providers in the metastore. Providers represent external organizations that share data with your metastore via Delta Sharing. Each provider can expose one or more shares. Returns: JSON array of provider objects, each containing name, comment, authentication_type, and recipient profile information. Results are capped at 100 items.

List all refresh runs for a quality monitor. Returns the history of refresh runs for the specified monitor, including completed, running, and failed refreshes. Args: table_name: Full three-level name of the monitored table in "catalog.schema.table" format. Returns: JSON array of refresh objects with refresh_id, state, start_time, and end_time. Results are capped at 100 items.

List saved SQL queries accessible to the current user. Returns queries ordered by creation time. Warning: calling this API concurrently 10+ times may result in throttling. Returns: JSON array of query objects with id, name, query_text, and creation metadata. Results are capped at 100 items.

List recent query execution history. Returns recently executed queries with their status, duration, and other execution metadata. Can optionally filter by warehouse. Args: warehouse_id: Optional warehouse ID to filter history. If empty, returns history across all warehouses. max_results: Maximum number of history entries to return (1-25, default 25). Returns: JSON array of query history entries with query_id, status, query_text, duration, and warehouse information.

List all Delta Sharing recipients in the metastore. Recipients represent external organizations or users who are authorized to consume shared data. Each recipient has an authentication type and associated credentials or tokens. Returns: JSON array of recipient objects, each containing name, authentication_type, comment, sharing_code (if TOKEN auth), and activation status. Results are capped at 100 items.

List registered models in Unity Catalog. Can optionally filter by catalog and schema. If both are empty, lists all registered models accessible to the caller. Args: catalog_name: Optional catalog name to filter models. If provided with schema_name, lists models in that schema. schema_name: Optional schema name to filter models. Requires catalog_name to also be provided. Returns: JSON array of registered model objects with name, catalog_name, schema_name, owner, and comment. Results are capped at 100 items.

List all Git repositories configured in the workspace. Returns repos accessible to the current user, including their URL, provider, branch, and workspace path. Returns: JSON array of repo objects with id, url, provider, path, branch, and head_commit_id. Results are capped at 100 items.

List job runs, optionally filtered by job ID. Returns runs in descending order by start time. Each run includes status, start/end times, task results, and error messages if any. Args: job_id: Filter runs to this specific job. If 0 or omitted, returns runs across all jobs. limit: Maximum number of runs to return. Must be between 1 and 100. Defaults to 25.

List all schemas within a catalog. Args: catalog_name: Name of the parent catalog to list schemas from. Returns: JSON array of schema objects with name, catalog_name, comment, and other metadata. Results are capped at 100 items.

List access control entries for a secret scope. Returns all ACLs that have been set on the given scope, showing which principals have READ, WRITE, or MANAGE permissions. Args: scope: Name of the secret scope to list ACLs for. Returns: JSON array of ACL objects, each containing principal and permission.

List all secret scopes in the workspace. Secret scopes are containers for secrets. Each scope has a name and a backend type (DATABRICKS or AZURE_KEYVAULT). Returns: JSON array of secret scope objects, each containing name, backend_type, and keyvault_metadata (if applicable).

List the metadata of all secrets within a scope. Returns key names and last-updated timestamps only. Secret values are never returned through the API for security reasons. Args: scope: Name of the secret scope to list secrets from. Returns: JSON array of secret metadata objects, each containing the key name and last_updated_timestamp. Does NOT include secret values.

List service principals in the workspace with optional SCIM filtering. Service principals are machine identities used for automated processes and CI/CD pipelines. Args: filter_str: Optional SCIM filter expression to narrow results. Examples: - 'displayName eq "my-pipeline"' - 'applicationId eq "00000000-0000-0000-0000-000000000000"' Leave empty to list all service principals. count: Maximum number of service principals to return (default 100). Returns: JSON array of service principal objects, each containing id, …

List permissions on a serving endpoint. Returns the access control list (ACL) for the specified endpoint, showing which users and groups have what level of access. Args: name: Name of the serving endpoint.

List all model serving endpoints in the workspace. Returns summary information for every serving endpoint including name, state, creation time, and served entity configuration.

List all Delta Sharing shares in the metastore. Shares are named objects that contain a collection of tables and schemas from one or more catalogs. They define what data is available for sharing with external recipients. Returns: JSON array of share objects, each containing name, owner, comment, and created/updated timestamps. Results are capped at 100 items.

List all available Spark runtime versions. Returns the Spark versions available for creating or editing clusters. Each entry includes the version key (used in cluster configuration) and a human-readable name. Returns: JSON object with available Spark versions and their display names.

List all storage credentials in the Unity Catalog metastore. Storage credentials define authentication methods for accessing cloud storage (AWS S3, Azure ADLS, GCS). Each credential can be used by one or more external locations. Returns: JSON array of storage credential objects with name, owner, cloud-specific credential info, and metadata. Results are capped at 100 items.

List all tables within a schema. Args: catalog_name: Name of the parent catalog. schema_name: Name of the schema to list tables from. Returns: JSON array of table objects with name, catalog_name, schema_name, table_type, and other metadata. Results are capped at 100 items.

List all personal access tokens in the workspace (admin only). Workspace admins can use this to audit all tokens across all users. Returns metadata about each token including owner, creation time, and expiry. Returns: JSON array of token info objects with token_id, owner, comment, creation_time, and expiry_time. Results are capped at 100 items.

List all personal access tokens for the current user. Returns metadata about each token including its ID, comment, creation time, and expiry. Token values are never returned. Returns: JSON array of token info objects with token_id, comment, creation_time, and expiry_time. Results are capped at 100 items.

List workspace users with optional SCIM filtering. Uses the SCIM 2.0 API to list users. Supports SCIM filter expressions for server-side filtering (e.g. 'displayName co "john"'). Args: filter_str: Optional SCIM filter expression to narrow results. Examples: - 'userName eq "user@example.com"' (exact match) - 'displayName co "john"' (contains) - 'active eq true' (only active users) Leave empty to list all users. count: Maximum number of users to return (default 100). Returns: JSON array of user objects…

List all vector search endpoints in the workspace. Vector search endpoints are the compute resources that host vector search indexes. Returns name, state, creation time, and endpoint type for each.

List all vector search indexes hosted on a specific endpoint. Returns summary information for each index including name, type, status, and the source table (for DELTA_SYNC indexes). Args: endpoint_name: Name of the vector search endpoint whose indexes to list.

List all volumes within a schema. Volumes are Unity Catalog objects that provide access to cloud storage locations for data files. Args: catalog_name: Name of the parent catalog. schema_name: Name of the schema to list volumes from. Returns: JSON array of volume objects with name, catalog_name, schema_name, volume_type, and storage_location. Results are capped at 100 items.

List all SQL warehouses in the workspace. Returns a JSON array of warehouse objects including id, name, state, cluster_size, and configuration details. Results are capped at 100 items.

List objects in a workspace directory. Returns notebooks, files, directories, and other objects at the specified path. Does not recurse into subdirectories. Args: path: Absolute workspace path to list (default: root "/"). Example: "/Users/user@example.com" or "/Repos". Returns: JSON array of workspace objects with path, object_type (NOTEBOOK, DIRECTORY, FILE, etc.), language, and size. Results are capped at 100 items.

Log a metric value for an MLflow run. Metrics are numeric measurements that track model performance over time. Multiple values can be logged for the same key at different steps to record training curves and convergence behavior. Args: run_id: The UUID of the run to log the metric to. key: Name of the metric (e.g. "rmse", "accuracy", "loss"). value: Numeric value of the metric at this step. step: Optional integer step number for tracking metrics over time. Defaults to 0. Use sequential step values to record training progress (e.g. epoch number or batch coun…

Log a parameter for an MLflow run. Parameters are key-value pairs that capture the configuration used for a training run (e.g. hyperparameters, data paths, model type). Each key can only be logged once per run. Args: run_id: The UUID of the run to log the parameter to. key: Name of the parameter (e.g. "learning_rate", "batch_size", "model_type"). value: String value of the parameter. Numeric values should be passed as strings (e.g. "0.001", "32"). Returns: Confirmation message on success.

Get a summary of the metastore assigned to the current workspace. Returns summary information about the metastore including its configuration and current status. Returns: JSON object with metastore summary including metastore_id, name, storage_root, owner, cloud, and global_metastore_id.

Migrate a classic SQL dashboard to Lakeview format. Converts an existing classic (legacy) SQL dashboard into a new Lakeview dashboard. The original dashboard is not modified. The migration creates a new Lakeview dashboard with the converted content. Args: source_dashboard_id: The UUID of the classic SQL dashboard to migrate. display_name: Optional display name for the new Lakeview dashboard. If empty, uses the original dashboard's name. parent_path: Optional workspace folder path for the new dashboard. Returns: JSON object with the …

Create a directory in the workspace. Creates the directory and all necessary parent directories if they do not exist (similar to "mkdir -p"). Args: path: Absolute workspace path of the directory to create. Example: "/Users/user@example.com/my_project/data". Returns: Confirmation message on success.

Pin a compute cluster to prevent it from being auto-deleted. Pinned clusters are retained even after being terminated for an extended period. This is useful for clusters that are used intermittently but should not be cleaned up. Args: cluster_id: The unique identifier of the cluster to pin. Returns: Confirmation that the cluster has been pinned.

Publish a Lakeview dashboard draft. Makes the current draft version of the dashboard publicly accessible (within the workspace). Published dashboards can be viewed by users who have access, and can use embedded credentials for query execution. Args: dashboard_id: The UUID identifying the dashboard to publish. warehouse_id: Optional warehouse ID to override the warehouse set in the draft. If empty, uses the draft's warehouse. embed_credentials: If True (default), the publisher's credentials are embedded so viewers can run queries without …

Store a secret value in a scope, creating or overwriting the key. If the key already exists, its value is overwritten. The caller must have WRITE or MANAGE permission on the scope. Args: scope: Name of the secret scope to store the secret in. key: Name of the secret key. Must be unique within the scope. string_value: The secret value to store as a UTF-8 string. Returns: Confirmation message on success.

Set an access control entry on a secret scope. Creates or overwrites the ACL for the given principal on the scope. The caller must have MANAGE permission on the scope. Args: scope: Name of the secret scope to set the ACL on. principal: The principal (user or group) to grant access to. Use the user's email address or the group name. permission: Permission level to grant. Must be one of: "READ" (can read secrets), "WRITE" (can read and write secrets), or "MANAGE" (full control including ACLs). Returns: Confirmation message on succ…

Execute a SQL query and return results formatted as a markdown table. Combines SQL execution with result formatting for easy display. Results are limited to 100 rows. Args: warehouse_id: ID of the SQL warehouse to use. sql: The SQL statement to execute. catalog: Optional catalog context for the query. schema: Optional schema context for the query.

Query a model serving endpoint for predictions. Sends input data to the serving endpoint and returns model predictions. Supports various input formats depending on the model type. Args: name: Name of the serving endpoint to query. inputs: JSON string containing the input data. The format depends on the model type: - For ML models: '[{"col1": val1, "col2": val2}, ...]' - For LLM completions: '{"prompt": "text"}' - For embeddings: '{"input": ["text1", "text2"]}' The string is parsed as JSON and passed as the 'inputs' parameter…

Query a vector search index for similar items. Performs approximate nearest neighbor (ANN) search to find vectors similar to the provided query. Provide either query_text (for indexes with a managed embedding model) or query_vector (for self-managed embeddings). Args: index_name: Full three-level name of the index to query. query_text: Text query string. Used with DELTA_SYNC indexes that have a configured embedding model endpoint. The text is automatically embedded before searching. query_vector: JSON array string of floats representing the query e…

Repair a failed job run by re-running specific tasks. Re-executes failed or skipped tasks within a completed run. This is useful for recovering from transient failures without re-running the entire job. Only tasks in a FAILED, TIMED_OUT, or CANCELED state can be repaired. Args: run_id: The unique numeric identifier of the run to repair. rerun_tasks: Comma-separated list of task keys to re-run (e.g. "task_1,task_2,task_3"). If empty, all failed tasks are re-run.

Resize a running cluster to a different number of workers. Changes the number of worker nodes in a running cluster. The resize happens gracefully — existing tasks are not interrupted. Args: cluster_id: The unique identifier of the cluster to resize. num_workers: The new number of worker nodes. Use 0 for a single-node cluster (driver only). Returns: Confirmation that the resize was initiated.

Restart a running compute cluster. Restarts the Spark driver and all workers. Use this to clear cached state or apply configuration changes. This call returns immediately without blocking. Args: cluster_id: The unique identifier of the cluster to restart. Returns: Confirmation that the restart was initiated.

Restore a previously deleted MLflow experiment. Recovers an experiment that was soft-deleted via databricks_delete_experiment. All runs within the experiment are also restored. Args: experiment_id: The numeric ID of the experiment to restore. Returns: Confirmation message on success.

Trigger a new run of an existing job. Starts the job immediately with its default parameters. The run executes asynchronously -- this tool returns the run ID without waiting for completion. Use databricks_get_run to check run status. Args: job_id: The unique numeric identifier of the job to run.

Trigger a refresh of the quality monitor for a table. Starts an asynchronous computation of data quality metrics. The refresh updates the profile and drift tables in the output schema. Args: table_name: Full three-level name of the monitored table in "catalog.schema.table" format. Returns: JSON object with refresh details including refresh_id and state.

Search for MLflow experiments using filter expressions. Supports filtering by experiment attributes and tags using the MLflow search syntax. Args: filter_string: Optional filter expression. Examples: - 'name = "my-experiment"' - 'tags.team = "ml-eng"' - 'attributes.lifecycle_stage = "active"' Leave empty to return all experiments. max_results: Maximum number of experiments to return (default 25). Returns: JSON array of matching experiment objects with experiment_id, name, artifact_lo…

Search for MLflow runs across one or more experiments. Supports filtering by metrics, parameters, tags, and run attributes using the MLflow search syntax. Args: experiment_ids: Comma-separated list of experiment IDs to search within (e.g. "123,456,789"). filter_string: Optional MLflow filter expression. Examples: - 'metrics.rmse < 0.5' - 'params.learning_rate = "0.01"' - 'tags.mlflow.source.type = "NOTEBOOK"' - 'attributes.status = "FINISHED"' Leave empty to return all r…

Set permissions on a workspace object. Replaces the entire access control list (ACL) for the specified object. All existing permissions not included in the new ACL will be removed. Args: object_type: The type of object to set permissions on. Common values: "clusters", "cluster-policies", "directories", "experiments", "jobs", "notebooks", "pipelines", "registered-models", "repos", "serving-endpoints", "sql/warehouses". object_id: The ID of the specific object. Format varies by object type. access_co…

Create a catalog (optionally) and schema in one step. Useful for quickly setting up a new data area. Creates the catalog if requested, then creates the schema within it. Args: catalog_name: Name of the parent catalog. schema_name: Name of the schema to create. comment: Optional description for the schema. create_catalog: If True, create the catalog first (skips if it already exists).

Start a Databricks App. Starts the last active deployment of the app. If the app is already running, this has no effect. This is an asynchronous operation. Use databricks_get_app to check when the app becomes active. Args: name: The name of the app to start. Returns: Confirmation that the app start has been initiated.

Start a terminated compute cluster. Initiates the start process for a previously terminated cluster. The cluster may take several minutes to become fully running. This call returns immediately without blocking. Args: cluster_id: The unique identifier of the cluster to start. Returns: Confirmation that the start was initiated.

Start a pipeline update (data refresh). Triggers the pipeline to process data. The update runs asynchronously -- this tool returns immediately without waiting for completion. Use databricks_get_pipeline to check the pipeline state. Args: pipeline_id: The unique identifier of the pipeline to start. full_refresh: If True, reprocesses all data from scratch, dropping and recreating all output tables. If False (default), performs an incremental update processing only new or changed data. Use full_refresh when schema changes have been m…

Start a stopped SQL warehouse. This initiates the start process. The warehouse may take several minutes to become fully running. This call returns immediately without waiting. Args: id: The unique identifier of the SQL warehouse to start. Returns: Confirmation that the start was initiated.

Stop a running Databricks App. Stops the active deployment of the app. The app's URL will become unavailable. Use databricks_start_app to restart it later. This is an asynchronous operation. Use databricks_get_app to confirm the app has stopped. Args: name: The name of the app to stop. Returns: Confirmation that the app stop has been initiated.

Stop a running pipeline. Sends a stop request to the pipeline. The pipeline finishes processing its current micro-batch (if any) and then stops. For continuous pipelines, this halts streaming. For triggered pipelines, this cancels the current update. Args: pipeline_id: The unique identifier of the pipeline to stop.

Stop a running SQL warehouse. Stops the warehouse to save costs. Any running queries will be terminated. This call returns immediately without waiting. Args: id: The unique identifier of the SQL warehouse to stop. Returns: Confirmation that the stop was initiated.

Trigger a sync of a DELTA_SYNC vector search index. Forces the index to re-sync with its source Delta table. This is useful when the source table has been updated and you want the index to reflect the latest data immediately rather than waiting for the next scheduled sync. Only applicable to DELTA_SYNC indexes. DIRECT_ACCESS indexes do not support this operation. Args: index_name: Full three-level name of the index (catalog.schema.index_name). Returns: Confirmation that the sync has been initiated.

Preview a table's schema and sample data in one call. Gets the table metadata (columns, types) and runs a SELECT query to show sample rows. If no warehouse_id is provided, finds a running warehouse automatically. Args: table_name: Full three-level table name (catalog.schema.table). warehouse_id: SQL warehouse ID to use. If empty, finds one automatically. limit: Number of sample rows to return. Defaults to 10.

Terminate a running compute cluster. Stops all Spark workloads and releases the cloud resources. The cluster configuration is preserved and it can be restarted later. This call returns immediately without blocking. Args: cluster_id: The unique identifier of the cluster to terminate. Returns: Confirmation that termination was initiated.

Find the right Databricks tools for a task or role. Call this tool when you're unsure which Databricks tools to use. Provide either a task description or a role name to get recommendations. Args: task: Describe what you want to do (e.g. "run a SQL query", "deploy an ML model", "create a user"). Leave empty to see all modules. role: One of: data_engineer, ml_engineer, platform_admin, app_developer, data_analyst, minimal. Returns the recommended module set for that role. Returns: JSON with recommended modules and their descriptions.

Move a Lakeview dashboard to the trash. Soft-deletes the dashboard by moving it to the trash. Trashed dashboards can potentially be recovered. The published version (if any) remains accessible until explicitly unpublished. Args: dashboard_id: The UUID identifying the dashboard to trash. Returns: Confirmation that the dashboard has been moved to the trash.

Unpin a compute cluster, allowing it to be auto-deleted. Removes the pin from a cluster so that it follows the normal auto-deletion policy for terminated clusters. Args: cluster_id: The unique identifier of the cluster to unpin. Returns: Confirmation that the cluster has been unpinned.

Unpublish a Lakeview dashboard. Removes the published version of the dashboard, making it no longer accessible to viewers. The draft version is retained and can be re-published later. Args: dashboard_id: The UUID identifying the dashboard to unpublish. Returns: Confirmation that the dashboard has been unpublished.

Update properties of an existing catalog. The caller must be the catalog owner or a metastore admin. Only non-empty fields are updated; other fields remain unchanged. Args: name: Name of the catalog to update. comment: Optional new description for the catalog. If empty, the comment is not changed. owner: Optional new owner for the catalog. If empty, the owner is not changed. Returns: JSON object with the updated catalog's details.

Update an existing connection's options or name. Updates the connection configuration. At least one of options_json or new_name must be provided. The caller must be the connection owner or a metastore admin. Args: name: Current name of the connection to update. options_json: Optional JSON string of key-value options to set. This replaces the existing options entirely, so include all desired options. Example: '{"host": "new-host.example.com", "port": "5432"}' new_name: Optional new name for the connection. Leave empty to …

Update an existing Lakeview dashboard draft. Updates the specified dashboard with new display name and/or layout. Only the provided non-empty fields will be updated; other fields remain unchanged. Args: dashboard_id: The UUID identifying the dashboard to update. display_name: New display name for the dashboard. If empty, the name is not changed. serialized_dashboard: New JSON string containing the updated dashboard layout. If empty, the layout is not changed. Returns: JSON object with the updated dashboard…

Update an existing Git credential. Replaces the provider, username, and token for the specified credential. All fields are required. Args: credential_id: The unique numeric identifier of the credential to update. git_provider: The Git provider name (e.g. "gitHub", "gitLab"). git_username: The new Git username. personal_access_token: The new personal access token. Stored securely and never returned in responses. Returns: Confirmation message on success.

Update an existing global init script. Replaces the script's name, content, enabled status, and position. All fields are required -- any omitted optional fields will be reset to their defaults. Args: script_id: The unique identifier of the script to update. name: New display name for the script. script: New script content, base64-encoded. enabled: Whether the script should be active after the update. position: New execution order position (0-based). Returns: Confirmation message on success.

Update (grant or revoke) permissions on a Unity Catalog securable. Applies a list of permission changes. Each change specifies a principal, a list of privileges to add, and a list of privileges to remove. Args: securable_type: The type of the securable object. Must be one of: "CATALOG", "SCHEMA", "TABLE", "VOLUME", "FUNCTION", "REGISTERED_MODEL", "EXTERNAL_LOCATION", "STORAGE_CREDENTIAL", "METASTORE", "CONNECTION", "SHARE", "RECIPIENT", "PROVIDER". full_name: The full name of the securable. Format depends…

Update the settings of an existing job. Replaces the job settings with the provided configuration. Use databricks_get_job to retrieve the current settings, modify them, and pass the updated JSON back. Args: job_id: The unique numeric identifier of the job to update. new_settings: JSON string containing the new job settings. Must be a valid JobSettings object serialized as JSON. Example: '{"name": "updated-job", "timeout_seconds": 3600}'

Update properties of an existing metastore. Only provided fields are updated. The caller must be the metastore owner or an account admin. Args: id: The unique identifier (UUID) of the metastore to update. name: Optional new name for the metastore. owner: Optional new owner of the metastore (user or group name). storage_root_credential_id: Optional ID of the storage credential to use for the metastore's root storage. Returns: JSON object with the updated metastore details.

Update an existing DLT pipeline configuration. Modifies the pipeline's metadata fields. Only non-empty parameters are updated; empty strings are ignored. To modify libraries or cluster config, use the Databricks UI or API directly. Note: The pipeline must be stopped before updating. If the pipeline is running, stop it first with databricks_stop_pipeline. Args: pipeline_id: The unique identifier of the pipeline to update. name: New name for the pipeline. Leave empty to keep the current name. target: New target schema. Leave empty to keep the current target. catalog: New Unity…

Update the quality monitor configuration for a table. Modifies the monitor's output schema, assets directory, slicing expressions, or baseline table. A refresh is needed after the update for changes to take effect on metrics. Args: table_name: Full three-level name of the monitored table in "catalog.schema.table" format. output_schema_name: Full name of the schema for profile and drift tables in "catalog.schema" format. assets_dir: Path for monitor dashboard and config assets. slicing_exprs: Optional comma-separated list of column expre…

Update a Git repository to a different branch or tag. Checks out the specified branch in the repository. This updates all files in the repo to match the branch's HEAD commit. Args: repo_id: The numeric ID of the repository. branch: The name of the branch to check out (e.g. "main", "develop"). Returns: JSON object with the updated repo details including the new branch and head_commit_id.

Update the status of an MLflow run. Typically used to mark a run as FINISHED, FAILED, or KILLED after training completes. Can also set the end time. Args: run_id: The UUID of the run to update. status: New status for the run. Valid values: "RUNNING", "SCHEDULED", "FINISHED", "FAILED", "KILLED". If empty, the status is not changed. end_time: Optional Unix timestamp in milliseconds for the run end time. If 0, the end time is not set. Returns: JSON object with the updated run info.

Update properties of an existing schema. The caller must be the schema owner or an owner of the parent catalog. Only non-empty fields are updated; other fields remain unchanged. Args: full_name: Full name of the schema in "catalog.schema" format (e.g. "my_catalog.my_schema"). comment: Optional new description for the schema. If empty, the comment is not changed. owner: Optional new owner for the schema. If empty, the owner is not changed. Returns: JSON object with the updated schema's details.

Update the configuration of an existing serving endpoint. Replaces the served entities with the specified model and version. An endpoint that already has an update in progress cannot be updated until the current update completes or fails. This tool returns immediately without waiting for the update to finish. Args: name: Name of the existing serving endpoint to update. model_name: Full name of the model to serve. model_version: Version of the model to serve. workload_size: Compute size per replica. One of "Small", "Medium", "Large". scale_to_zero: Whether the endpoint sca…

Update properties of an existing Delta Sharing share. The caller must be the share owner or a metastore admin. Only non-empty fields are updated; other fields remain unchanged. Args: name: The name of the share to update. comment: Optional new description for the share. If empty, the comment is not changed. owner: Optional new owner for the share. If empty, the owner is not changed. Returns: JSON object with the updated share's details.

Update an existing storage credential. Allows renaming, changing the comment, or updating the cloud-specific authentication configuration. Only provided fields are updated. Args: name: Current name of the storage credential to update. new_name: Optional new name for the credential. comment: Optional new description. aws_iam_role_arn: For AWS — updated IAM role ARN. Returns: JSON object with the updated storage credential details.

Validate a storage credential's access to a cloud storage location. Tests whether the credential can access the specified URL. Can validate an existing credential by name or test a new credential configuration before creating it. Args: storage_credential_name: Name of an existing credential to validate. If empty, provide aws_iam_role_arn to test a new configuration. url: Cloud storage URL to validate access against (e.g. "s3://my-bucket/path" or "abfss://container@account.dfs.core.windows.net/path"). aws_iam_role_…

Get a comprehensive status overview of the Databricks workspace. Returns a health dashboard showing the status of clusters, warehouses, jobs, serving endpoints, and catalog access -- all in one call.