dbt Cloud Run Completion to Metadata Discovery

arazzo: 1.0.1
info:
  title: dbt Cloud Run Completion to Metadata Discovery
  summary: Confirm a run succeeded, then query the Discovery API for the models it produced.
  description: >-
    Bridges the Administrative API and the Discovery API. The workflow reads a
    run and confirms it reached the terminal success status (status code 10),
    then issues a GraphQL query to the Discovery API to pull model metadata for
    the run's environment. This is the standard pattern for hydrating a metadata
    catalog after a successful dbt build. 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: metadataApi
  url: ../openapi/dbt-metadata-api-openapi.yml
  type: openapi
- name: runsApi
  url: ../openapi/dbt-runs-api-openapi.yml
  type: openapi
workflows:
- workflowId: run-metadata-discovery
  summary: After a successful run, query the Discovery API for its environment's models.
  description: >-
    Reads the run; if status is 10 (success) it runs a Discovery API GraphQL
    query for the run's environment id to retrieve produced model metadata.
  inputs:
    type: object
    required:
    - accountId
    - runId
    - token
    properties:
      accountId:
        type: integer
        description: The dbt Cloud account id.
      runId:
        type: integer
        description: The id of the run to confirm and trace metadata for.
      token:
        type: string
        description: The dbt Cloud service token (sent as Authorization Bearer Token).
  steps:
  - stepId: getRun
    description: >-
      Read the run and confirm it succeeded. Status 10 indicates the run
      completed successfully and its metadata is available in the Discovery API.
    operationId: getRun
    parameters:
    - name: Authorization
      in: header
      value: Bearer $inputs.token
    - name: accountId
      in: path
      value: $inputs.accountId
    - name: runId
      in: path
      value: $inputs.runId
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      status: $response.body#/data/status
      environmentId: $response.body#/data/environment_id
    onSuccess:
    - name: succeeded
      type: goto
      stepId: queryMetadata
      criteria:
      - context: $response.body
        condition: $.data.status == 10
        type: jsonpath
    - name: notSucceeded
      type: end
      criteria:
      - context: $response.body
        condition: $.data.status != 10
        type: jsonpath
  - stepId: queryMetadata
    description: >-
      Query the Discovery API for the models in the run's environment using a
      GraphQL query whose environmentId variable is bound to the run's
      environment id.
    operationId: executeDiscoveryQuery
    parameters:
    - name: Authorization
      in: header
      value: Bearer $inputs.token
    requestBody:
      contentType: application/json
      payload:
        query: >-
          query Models($environmentId: BigInt!) {
            environment(id: $environmentId) {
              applied {
                models(first: 100) { edges { node { name uniqueId materializedType } } }
              }
            }
          }
        variables:
          environmentId: $steps.getRun.outputs.environmentId
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      models: $response.body#/data/environment/applied/models
  outputs:
    runStatus: $steps.getRun.outputs.status
    environmentId: $steps.getRun.outputs.environmentId
    models: $steps.queryMetadata.outputs.models