arazzo: 1.0.1
info:
  title: Grafana Export Alerting as Code
  summary: Export contact points, the notification policy tree, and alert rules as provisioning files.
  description: >-
    Most Grafana alerting is clicked into existence and then lives nowhere but
    the database. This workflow pulls the three halves of an alerting setup —
    contact points, the notification policy tree, and the alert rules — out in
    Grafana's own provisioning file format, so a running instance can be committed
    to a repository and the drift between what is deployed and what is reviewed
    becomes visible. Run it on a schedule and diff the output. 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: grafanaApi
  url: ../openapi/grafana-openapi.yml
  type: openapi
workflows:
- workflowId: export-alerting-as-code
  summary: Snapshot a Grafana instance's alerting configuration into provisioning YAML.
  description: >-
    Exports contact points, then the routing tree, then the alert rules, each in
    provisioning file format ready to commit.
  inputs:
    type: object
    properties:
      format:
        type: string
        description: Export format; yaml, json, or hcl.
        default: yaml
      folderUid:
        type: string
        description: >-
          Limit the alert rule export to a single folder. Omit to export every
          rule the caller can see.
      group:
        type: string
        description: >-
          Limit the alert rule export to one evaluation group within the folder.
          Requires folderUid.
      decrypt:
        type: boolean
        description: >-
          Include contact point secrets in plaintext. Requires admin permission,
          and anything exported this way must be treated as a secret — leave it
          false unless the destination is a secret store.
        default: false
  steps:
  - stepId: exportContactPoints
    description: >-
      Export the contact points. Secrets are redacted unless decrypt is set, so
      the default output is safe to commit to a repository.
    operationId: routeGetContactpointsExport
    parameters:
    - name: format
      in: query
      value: $inputs.format
    - name: decrypt
      in: query
      value: $inputs.decrypt
    - name: download
      in: query
      value: false
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      contactPoints: $response.body
  - stepId: exportPolicyTree
    description: >-
      Export the notification policy tree. This is the routing that binds alert
      labels to the contact points exported above, and it is meaningless without
      them, which is why the two are captured in the same run.
    operationId: routeGetPolicyTreeExport
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      policyTree: $response.body
  - stepId: exportAlertRules
    description: >-
      Export the alert rules, optionally scoped to one folder or evaluation
      group, in provisioning file format.
    operationId: routeGetAlertRulesExport
    parameters:
    - name: format
      in: query
      value: $inputs.format
    - name: folderUid
      in: query
      value: $inputs.folderUid
    - name: group
      in: query
      value: $inputs.group
    - name: download
      in: query
      value: false
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      alertRules: $response.body
  outputs:
    contactPoints: $steps.exportContactPoints.outputs.contactPoints
    policyTree: $steps.exportPolicyTree.outputs.policyTree
    alertRules: $steps.exportAlertRules.outputs.alertRules