Azure Databricks Preflight and Create a Cluster

arazzo: 1.0.1
info:
  title: Azure Databricks Preflight and Create a Cluster
  summary: Resolve a valid Spark version and node type, then create a cluster.
  description: >-
    Discovers valid cluster configuration values before provisioning. The
    workflow lists the available Databricks Runtime versions and supported node
    types, selects the first of each as defaults, and creates a cluster using
    those resolved values. This avoids hard-coding version or node identifiers
    that may not exist in the target workspace. Every step spells out its
    request inline so the flow can be read and executed without opening the
    underlying OpenAPI description.
  version: 1.0.0
sourceDescriptions:
- name: azureDatabricksApi
  url: ../openapi/azure-databricks-openapi.yml
  type: openapi
workflows:
- workflowId: preflight-create-cluster
  summary: Resolve a Spark version and node type, then create a cluster.
  description: >-
    Lists spark versions and node types, picks the first of each, then calls
    createCluster with those resolved values.
  inputs:
    type: object
    required:
    - token
    - clusterName
    - numWorkers
    properties:
      token:
        type: string
        description: Databricks personal access token for the Authorization header.
      clusterName:
        type: string
        description: Name for the new cluster.
      numWorkers:
        type: integer
        description: Number of worker nodes for the cluster.
  steps:
  - stepId: listSparkVersions
    description: >-
      List the available Databricks Runtime versions and capture the first
      version key to use for the cluster.
    operationId: listSparkVersions
    parameters:
    - name: Authorization
      in: header
      value: Bearer $inputs.token
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      sparkVersion: $response.body#/versions/0/key
  - stepId: listNodeTypes
    description: >-
      List the supported Azure VM node types and capture the first node type id
      to use for the cluster.
    operationId: listNodeTypes
    parameters:
    - name: Authorization
      in: header
      value: Bearer $inputs.token
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      nodeTypeId: $response.body#/node_types/0/node_type_id
  - stepId: createCluster
    description: >-
      Create the cluster using the resolved Spark version and node type,
      capturing the returned cluster_id.
    operationId: createCluster
    parameters:
    - name: Authorization
      in: header
      value: Bearer $inputs.token
    requestBody:
      contentType: application/json
      payload:
        cluster_name: $inputs.clusterName
        spark_version: $steps.listSparkVersions.outputs.sparkVersion
        node_type_id: $steps.listNodeTypes.outputs.nodeTypeId
        num_workers: $inputs.numWorkers
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      clusterId: $response.body#/cluster_id
  outputs:
    clusterId: $steps.createCluster.outputs.clusterId
    sparkVersion: $steps.listSparkVersions.outputs.sparkVersion
    nodeTypeId: $steps.listNodeTypes.outputs.nodeTypeId