arazzo: 1.0.1
info:
  title: Azure Synapse Analytics Submit and Poll Spark Batch Job
  summary: Submit a Spark batch job to a pool and poll until it reaches a result.
  description: >-
    Spark batch jobs run application code against a Synapse Spark pool through
    the Livy API. This workflow submits a batch job, then polls the job by id in
    a loop, branching back to poll again while the job is still running and
    ending once a terminal result (Succeeded, Failed, or Cancelled) is reported.
    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: sparkJobApi
  url: ../openapi/azure-synapse-analytics-spark-job-openapi.yml
  type: openapi
workflows:
- workflowId: submit-and-poll-spark-batch-job
  summary: Submit a Spark batch job and poll it until it reaches a terminal result.
  description: >-
    Creates a Spark batch job against a pool, reads the job by id, and loops on
    the read until the livy result is no longer Uncertain, capturing the final
    result.
  inputs:
    type: object
    required:
    - sparkPoolName
    - batchOptions
    properties:
      sparkPoolName:
        type: string
        description: The name of the Spark pool to run the batch job on.
      batchOptions:
        type: object
        description: >-
          The SparkBatchJobOptions body. Must include name and file at minimum.
  steps:
  - stepId: submitBatchJob
    description: >-
      Submit a new Spark batch job to the pool using the supplied batch options.
    operationId: SparkBatch_CreateSparkBatchJob
    parameters:
    - name: sparkPoolName
      in: path
      value: $inputs.sparkPoolName
    - name: detailed
      in: query
      value: true
    requestBody:
      contentType: application/json
      payload: $inputs.batchOptions
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      batchId: $response.body#/id
      submittedState: $response.body#/state
  - stepId: pollBatchJob
    description: >-
      Read the batch job by id. While the livy result is still Uncertain, loop
      back and poll again; once a terminal result is reported, end the workflow.
    operationId: SparkBatch_GetSparkBatchJob
    parameters:
    - name: sparkPoolName
      in: path
      value: $inputs.sparkPoolName
    - name: batchId
      in: path
      value: $steps.submitBatchJob.outputs.batchId
    - name: detailed
      in: query
      value: true
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      result: $response.body#/result
      currentState: $response.body#/state
    onSuccess:
    - name: stillRunning
      type: goto
      stepId: pollBatchJob
      criteria:
      - context: $response.body
        condition: $.result == "Uncertain"
        type: jsonpath
    - name: finished
      type: end
      criteria:
      - context: $response.body
        condition: $.result != "Uncertain"
        type: jsonpath
  outputs:
    batchId: $steps.submitBatchJob.outputs.batchId
    result: $steps.pollBatchJob.outputs.result