arazzo: 1.0.1
info:
  title: Dynatrace Query a Metric and Raise an Alert Event
  summary: Confirm a metric exists, query its data points, and raise a custom alert event.
  description: >-
    A monitoring-as-code loop that turns a metric reading into a Dynatrace alert
    annotation. The workflow first reads the metric descriptor to confirm the
    metric and its supported aggregations, queries recent data points for the
    target entities, and then ingests a CUSTOM_ALERT event describing the
    condition. 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: metricsApi
  url: ../openapi/dynatrace-metrics-api-v2-openapi.yml
  type: openapi
- name: eventsApi
  url: ../openapi/dynatrace-events-api-v2-openapi.yml
  type: openapi
workflows:
- workflowId: metric-threshold-alert
  summary: Read a metric, query its data, and raise a custom alert event.
  description: >-
    Confirms a metric descriptor exists, queries its data points over a time
    range scoped to an entity selector, and ingests a custom alert event.
  inputs:
    type: object
    required:
    - metricKey
    - metricSelector
    - entitySelector
    - alertTitle
    properties:
      metricKey:
        type: string
        description: The metric key to inspect (e.g. builtin:host.cpu.usage).
      metricSelector:
        type: string
        description: The metric selector with transformations (e.g. builtin:host.cpu.usage:avg).
      entitySelector:
        type: string
        description: The entity selector scoping the query (e.g. type(HOST),tag(prod)).
      resolution:
        type: string
        description: The desired data-point resolution (e.g. 5m, 1h).
      from:
        type: string
        description: Start of the queried time range (e.g. now-2h).
      alertTitle:
        type: string
        description: The title of the custom alert event to raise.
  steps:
  - stepId: getDescriptor
    description: >-
      Read the metric descriptor to confirm the metric exists and to learn its
      unit and supported aggregation types before querying data.
    operationId: getMetricDescriptor
    parameters:
    - name: metricKey
      in: path
      value: $inputs.metricKey
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      displayName: $response.body#/displayName
      unit: $response.body#/unit
  - stepId: queryData
    description: >-
      Query data points for the metric selector over the time range, scoped to
      the supplied entity selector and resolution.
    operationId: queryMetricData
    parameters:
    - name: metricSelector
      in: query
      value: $inputs.metricSelector
    - name: entitySelector
      in: query
      value: $inputs.entitySelector
    - name: resolution
      in: query
      value: $inputs.resolution
    - name: from
      in: query
      value: $inputs.from
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      resolution: $response.body#/resolution
      totalCount: $response.body#/totalCount
      firstMetricId: $response.body#/result/0/metricId
  - stepId: raiseAlertEvent
    description: >-
      Ingest a CUSTOM_ALERT event onto the affected entities' timeline,
      attaching the metric key and reading as properties.
    operationId: ingestEvent
    parameters: []
    requestBody:
      contentType: application/json
      payload:
        eventType: CUSTOM_ALERT
        title: $inputs.alertTitle
        entitySelector: $inputs.entitySelector
        properties:
          metricKey: $inputs.metricKey
          metricUnit: $steps.getDescriptor.outputs.unit
    successCriteria:
    - condition: $statusCode == 201
    outputs:
      reportCount: $response.body#/reportCount
      eventId: $response.body#/eventIngestResults/0/eventId
  outputs:
    metricDisplayName: $steps.getDescriptor.outputs.displayName
    seriesCount: $steps.queryData.outputs.totalCount
    eventId: $steps.raiseAlertEvent.outputs.eventId