Skip to main content

Databricks Component

Manage compute, workflow jobs, ML models, SQL queries and more within a Databricks workspace.

Component key: databricks

Description

Databricks is an analytics and artificial intelligence platform where you can build, scale and govern data and AI, including generative AI and other machine learning models. This component lets you interact with the Databricks REST API to manage clusters, jobs, libraries, and other resources.

API Documentation

This component was built using the Databricks REST API Reference.

Connections

Databricks Personal Access Token

While service principal authentication is the recommended method for authenticating with the Databricks REST API, you can also use personal access tokens (which are tied to specific users). To generate a personal access token:

  1. Open https://accounts.cloud.databricks.com/workspaces and select your workspace. Open the URL for your workspace (e.g., https://dbc-00000000-aaaa.cloud.databricks.com) and log in.
  2. From the top-right click your user icon and select Settings.
  3. Under the User > Developer tab, select Manage under Access tokens.
  4. Click the Generate New Token button. Enter a description for the token and click Generate. Omit Lifetime (days) to create a token that never expires.

Your token will look similar to dap000000000000000000000000000000000. Copy this token and use it as the token input for the Databricks components.

When configuring an instance in Prismatic, enter the personal access token along with your workspace's endpoint (like dbc-REPLACE-ME.cloud.databricks.com).

See https://docs.databricks.com/en/dev-tools/auth/pat.html for more information on personal access token authentication.

Databricks Workspace Service Principal

With service principal authentication, you create a service user within your account, grant the user permissions to a workspace, and then generate a client ID and secret pair for that service user. This component uses that key pair to authenticate with workspaces that the service account has been granted permissions to. This is the best practice for authenticating with the Databricks REST API.

To set up service principal authentication, you need to:

  1. Create the service principal
    1. Open https://accounts.cloud.databricks.com/users. Under the Service principals tab select Add service principal.
    2. Give the service principal any name and click Add.
  2. Grant the service principal permission to your workspace
    1. Navigate to https://accounts.cloud.databricks.com/workspaces and select your workspace
    2. Under the Permissions tab select Add permissions
    3. Search for the service principal you created and grant the permission Admin.
  3. Generate a key pair for the service principal
    1. Navigate to the service principal and open the Principal information tab.
    2. Under OAuth secrets select Generate secret.
    3. Take note of the Secret (i.e. "Client Secret") and Client ID you receive. The client ID should be a UUID like 00000000-0000-0000-0000-000000000000. The client secret will look like dose00000000000000000000000000000000.

When configuring an instance in Prismatic, enter the client ID and client secret along with your workspace's token URL endpoint (like https://dbc-REPLACE-ME.cloud.databricks.com/oidc/v1/token) if you need workspace-level API access.

If you need account-level access to the API (i.e. you need to manage workspaces using this service principal), you will need to grant the service principal administrative access to your account, and your token URL will look like https://accounts.cloud.databricks.com/oidc/accounts/<my-account-id>/v1/token.

See https://docs.databricks.com/en/dev-tools/auth/oauth-m2m.html for more information on service principal OAuth client credential authentication.

Data Sources

Select Cluster

Select a Databricks cluster to use | key: selectCluster | type: picklist


Select Node Type

Select a Databricks node type to use | key: selectNodeType | type: picklist


Select SQL Warehouse

Select an SQL Warehouse | key: selectWarehouse | type: picklist


Actions

Create Execution Context

Create a Databricks execution context | key: createExecutionContext

{
"data": {
"id": "1234-567890-reef123"
}
}

Get Cluster

Get a Databricks cluster by ID | key: getCluster

{
"data": {
"cluster_id": "1234-567890-reef123",
"spark_context_id": 4020997813441462000,
"cluster_name": "my-cluster",
"spark_version": "13.3.x-scala2.12",
"aws_attributes": {
"zone_id": "us-west-2c",
"first_on_demand": 1,
"availability": "SPOT_WITH_FALLBACK",
"spot_bid_price_percent": 100,
"ebs_volume_count": 0
},
"node_type_id": "i3.xlarge",
"driver_node_type_id": "i3.xlarge",
"autotermination_minutes": 120,
"enable_elastic_disk": false,
"disk_spec": {
"disk_count": 0
},
"cluster_source": "UI",
"enable_local_disk_encryption": false,
"instance_source": {
"node_type_id": "i3.xlarge"
},
"driver_instance_source": {
"node_type_id": "i3.xlarge"
},
"state": "TERMINATED",
"state_message": "Inactive cluster terminated (inactive for 120 minutes).",
"start_time": 1618263108824,
"terminated_time": 1619746525713,
"last_state_loss_time": 1619739324740,
"num_workers": 30,
"default_tags": {
"Vendor": "Databricks",
"Creator": "someone@example.com",
"ClusterName": "my-cluster",
"ClusterId": "1234-567890-reef123"
},
"creator_user_name": "someone@example.com",
"termination_reason": {
"code": "INACTIVITY",
"parameters": {
"inactivity_duration_min": "120"
},
"type": "SUCCESS"
},
"init_scripts_safe_mode": false,
"spec": {
"spark_version": "13.3.x-scala2.12"
}
}
}

Get Command Status

Gets the status of and, if available, the results from a currently executing command. | key: getCommandStatus

{
"data": {
"id": "d4aa2c2f871048e797efdbe635de94be",
"status": "Running",
"result": null
}
}

Get Current User

Get the currently authenticated Databricks user or service principal. | key: getCurrentUser

{
"data": {
"emails": [
{
"type": "work",
"value": "1d021345-e23c-4f29-84fa-d027a622259e",
"primary": true
}
],
"displayName": "Example Service User",
"schemas": [
"urn:ietf:params:scim:schemas:core:2.0:User",
"urn:ietf:params:scim:schemas:extension:workspace:2.0:User"
],
"name": {
"familyName": "User",
"givenName": "Example Service"
},
"active": true,
"groups": [
{
"display": "admins",
"type": "direct",
"value": "272831250941646",
"$ref": "Groups/272831250941646"
}
],
"id": "7556761598142352",
"userName": "1d021345-e23c-4f29-84fa-d027a622259e"
}
}

Get SQL Warehouse

Get an SQL Warehouse | key: getWarehouse

{
"data": {
"id": "0000000000000000",
"name": "Starter Warehouse",
"size": "SMALL",
"cluster_size": "Small",
"min_num_clusters": 1,
"max_num_clusters": 1,
"auto_stop_mins": 60,
"auto_resume": true,
"creator_name": "example@example.com",
"creator_id": 5760885597616698,
"tags": {},
"spot_instance_policy": "COST_OPTIMIZED",
"enable_photon": true,
"enable_serverless_compute": false,
"warehouse_type": "PRO",
"num_clusters": 1,
"num_active_sessions": 0,
"state": "RUNNING",
"jdbc_url": "jdbc:spark://dbc-example-0000.cloud.databricks.com:443/default;transportMode=http;ssl=1;AuthMech=3;httpPath=/sql/1.0/warehouses/0000000000000000;",
"odbc_params": {
"hostname": "dbc-example-0000.cloud.databricks.com",
"path": "/sql/1.0/warehouses/0000000000000000",
"protocol": "https",
"port": 443
},
"health": {
"status": "HEALTHY"
}
}
}

List Clusters

Return information about all pinned clusters, active clusters, up to 200 of the most recently terminated all-purpose clusters in the past 30 days, and up to 30 of the most recently terminated job clusters in the past 30 days. | key: listClusters

{
"data": [
{
"cluster_id": "1234-567890-reef123",
"spark_context_id": 4020997813441462000,
"cluster_name": "my-cluster",
"spark_version": "13.3.x-scala2.12",
"aws_attributes": {
"zone_id": "us-west-2c",
"first_on_demand": 1,
"availability": "SPOT_WITH_FALLBACK",
"spot_bid_price_percent": 100,
"ebs_volume_count": 0
},
"node_type_id": "i3.xlarge",
"driver_node_type_id": "i3.xlarge",
"autotermination_minutes": 120,
"enable_elastic_disk": false,
"disk_spec": {
"disk_count": 0
},
"cluster_source": "UI",
"enable_local_disk_encryption": false,
"instance_source": {
"node_type_id": "i3.xlarge"
},
"driver_instance_source": {
"node_type_id": "i3.xlarge"
},
"state": "TERMINATED",
"state_message": "Inactive cluster terminated (inactive for 120 minutes).",
"start_time": 1618263108824,
"terminated_time": 1619746525713,
"last_state_loss_time": 1619739324740,
"num_workers": 30,
"default_tags": {
"Vendor": "Databricks",
"Creator": "someone@example.com",
"ClusterName": "my-cluster",
"ClusterId": "1234-567890-reef123"
},
"creator_user_name": "someone@example.com",
"termination_reason": {
"code": "INACTIVITY",
"parameters": {
"inactivity_duration_min": "120"
},
"type": "SUCCESS"
},
"init_scripts_safe_mode": false,
"spec": {
"spark_version": "13.3.x-scala2.12"
}
}
]
}

List Node Types

Returns a list of supported Spark node types. These node types can be used to launch a cluster. | key: listNodeTypes

{
"data": [
{
"node_type_id": "r4.xlarge",
"memory_mb": 31232,
"num_cores": 4,
"description": "r4.xlarge",
"instance_type_id": "r4.xlarge",
"is_deprecated": false,
"category": "Memory Optimized",
"support_ebs_volumes": true,
"support_cluster_tags": true,
"num_gpus": 0,
"node_instance_type": {
"instance_type_id": "r4.xlarge",
"local_disks": 0,
"local_disk_size_gb": 0,
"instance_family": "EC2 r4 Family vCPUs",
"swap_size": "10g"
},
"is_hidden": false,
"support_port_forwarding": true,
"supports_elastic_disk": true,
"display_order": 0,
"is_io_cache_enabled": false
}
]
}

List SQL Warehouses

List all SQL Warehouses in the Databricks workspace | key: listWarehouses

{
"data": [
{
"id": "0000000000000000",
"name": "Starter Warehouse",
"size": "SMALL",
"cluster_size": "Small",
"min_num_clusters": 1,
"max_num_clusters": 1,
"auto_stop_mins": 60,
"auto_resume": true,
"creator_name": "example@example.com",
"creator_id": 5760885597616698,
"tags": {},
"spot_instance_policy": "COST_OPTIMIZED",
"enable_photon": true,
"enable_serverless_compute": false,
"warehouse_type": "PRO",
"num_clusters": 1,
"num_active_sessions": 0,
"state": "RUNNING",
"jdbc_url": "jdbc:spark://dbc-example-0000.cloud.databricks.com:443/default;transportMode=http;ssl=1;AuthMech=3;httpPath=/sql/1.0/warehouses/0000000000000000;",
"odbc_params": {
"hostname": "dbc-example-0000.cloud.databricks.com",
"path": "/sql/1.0/warehouses/0000000000000000",
"protocol": "https",
"port": 443
},
"health": {
"status": "HEALTHY"
}
}
]
}

Raw Request

Send raw HTTP request to the Databricks API. | key: rawRequest


Restart Cluster

Restart a Databricks cluster by ID | key: restartCluster

{
"data": "Cluster restarted successfully"
}

Run Command

Run a command in a Databricks execution context | key: runCommand

{
"data": {
"id": "d4aa2c2f871048e797efdbe635de94be"
}
}

SQL: Execute an SQL Statement

Run a SQL query in the Databricks workspace. You can choose to wait for the result or asynchronously issue the request and return the statement ID. | key: runSql

Parameters can be passed to an SQL query using colon notation. Your parameters should be an array of objects, and each object should have a name and value property, and optional type property.

For example, if your statement reads SELECT * FROM my_table WHERE name = :my_name AND date = :my_date, you can pass the following parameters:

[
{ "name": "my_name", "value": "the name" },
{ "name": "my_date", "value": "2020-01-01", "type": "DATE" }
]

This action will execute the SQL query and then wait for the results to be available, throwing an error if the query fails.

If you expect your query to run for a long time, use the "Raw Request" action to issue a query, take note of the statement_id that is returned, and then use that statement ID to fetch results at a later time. Large results may be split into chunks, and you can use the the next_chunk_internal_link to fetch the next chunk of results. See https://docs.databricks.com/api/workspace/statementexecution/executestatement for more information.


Start SQL Warehouse

Start an SQL Warehouse | key: startWarehouse

{
"data": "Warehouse started"
}

Start Terminated Cluster

Start a terminated Databricks cluster by ID | key: startTerminatedCluster

{
"data": "Cluster started successfully"
}

Stop SQL Warehouse

Stop an SQL Warehouse | key: stopWarehouse

{
"data": "Warehouse stopped"
}

Terminate Cluster

Terminate a Databricks cluster by ID | key: terminateCluster

{
"data": "Cluster terminated successfully"
}