Databricks Provision Cluster and Wait Until Running

arazzo: 1.0.1
info:
  title: Databricks Provision Cluster and Wait Until Running
  summary: Create a Spark cluster and poll its state until it reaches RUNNING.
  description: >-
    Provisions a new Databricks Spark cluster from a supplied configuration and
    then polls the cluster details endpoint until the asynchronously created
    cluster transitions out of PENDING into the RUNNING state. Cluster creation
    is asynchronous, so the cluster_id returned by the create call is reused to
    drive the poll loop. 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
  x-realizes-capability-ids:
  - BC-600.50
  x-capability-derivation:
    method: 'deterministic join: sourceDescriptions -> per-tag OpenAPI -> tag/capability edge. No classification at this step.'
    min_confidence: 0.7
    sources:
    - capability_id: BC-600.50
      capability_name: IT Infrastructure Management
      spec: databricks-clusters-api-openapi.yml
      confidence: 0.75
    model: Turbo EA Capabilities by Vincent Verdet — Turbo EA, https://github.com/vincentmakes/turbo-ea-capabilities, CC BY 4.0
sourceDescriptions:
- name: clustersApi
  url: ../openapi/databricks-clusters-api-openapi.yml
  type: openapi
workflows:
- workflowId: provision-cluster
  summary: Create a cluster and wait until it is RUNNING.
  description: >-
    Submits a cluster creation request and then repeatedly reads the cluster
    status until the life cycle state becomes RUNNING, branching back to the
    poll step while the cluster is still PENDING.
  inputs:
    type: object
    required:
    - cluster_name
    - spark_version
    - node_type_id
    - num_workers
    properties:
      cluster_name:
        type: string
        description: Human-readable name for the new cluster.
      spark_version:
        type: string
        description: The Spark runtime version (e.g. a Databricks Runtime key).
      node_type_id:
        type: string
        description: The node type for worker and driver nodes.
      num_workers:
        type: integer
        description: The number of worker nodes for the cluster.
      autotermination_minutes:
        type: integer
        description: Idle minutes before the cluster auto-terminates.
  steps:
  - stepId: createCluster
    description: >-
      Create the Spark cluster. The response returns the cluster_id used to
      poll for readiness; creation itself is asynchronous.
    operationId: createCluster
    requestBody:
      contentType: application/json
      payload:
        cluster_name: $inputs.cluster_name
        spark_version: $inputs.spark_version
        node_type_id: $inputs.node_type_id
        num_workers: $inputs.num_workers
        autotermination_minutes: $inputs.autotermination_minutes
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      clusterId: $response.body#/cluster_id
  - stepId: pollClusterState
    description: >-
      Read the cluster details and inspect the life cycle state. Loop back
      while the cluster is still PENDING; finish once it is RUNNING.
    operationId: getCluster
    parameters:
    - name: cluster_id
      in: query
      value: $steps.createCluster.outputs.clusterId
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      state: $response.body#/state
      stateMessage: $response.body#/state_message
    onSuccess:
    - name: stillPending
      type: goto
      stepId: pollClusterState
      criteria:
      - context: $response.body
        condition: $.state == "PENDING"
        type: jsonpath
    - name: clusterRunning
      type: end
      criteria:
      - context: $response.body
        condition: $.state == "RUNNING"
        type: jsonpath
  outputs:
    clusterId: $steps.createCluster.outputs.clusterId
    finalState: $steps.pollClusterState.outputs.state