Prometheus Reload Configuration and Verify It Applied

arazzo: 1.0.1
info:
  title: Prometheus Reload Configuration and Verify It Applied
  summary: Trigger a config reload and prove the new configuration and rules are live.
  description: >-
    A configuration reload is the most routine change made to a running
    Prometheus, and the one most often assumed to have worked. The reload
    endpoint returning 200 only means the file parsed. The workflow captures the
    configuration before the change, triggers the reload, checks the runtime
    info where reloadConfigSuccess reports whether the server actually adopted
    it, re-reads the configuration to confirm the new content is live, and lists
    the rules to catch rule files that failed to load. 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: lifecycleApi
  url: ../openapi/prometheus-lifecycle-api-openapi.yml
  type: openapi
- name: rulesApi
  url: ../openapi/prometheus-rules-api-openapi.yml
  type: openapi
- name: statusApi
  url: ../openapi/prometheus-status-api-openapi.yml
  type: openapi
workflows:
- workflowId: reload-config-verify
  summary: Reload the configuration and confirm it took effect.
  description: >-
    The reload endpoint requires the --web.enable-lifecycle flag; without it the
    server will not expose it at all. reloadConfigSuccess in the runtime info is
    the authoritative signal that the reload was adopted, not just accepted.
  inputs:
    type: object
    properties:
      ruleType:
        type: string
        description: >-
          Optional rule type filter for the post-reload rule listing — alert or
          record. Omit to list both.
  steps:
  - stepId: captureConfigBefore
    description: >-
      Read the currently loaded configuration so there is a before image to
      compare against, and a record of what to restore if the reload goes wrong.
    operationId: getConfig
    successCriteria:
    - condition: $statusCode == 200
    - condition: $response.body#/status == "success"
    outputs:
      configYamlBefore: $response.body#/data/yaml
  - stepId: reloadConfig
    description: >-
      Trigger the reload of the configuration file and all rule files. A 500
      here means the new configuration failed to parse and the server has kept
      running on the old one.
    operationId: reloadConfig
    successCriteria:
    - condition: $statusCode == 200
  - stepId: confirmReloadSucceeded
    description: >-
      Read the runtime info and check reloadConfigSuccess. This is the field
      that separates a reload the server accepted from one it actually adopted,
      and lastConfigTime dates the adoption.
    operationId: getRuntimeInfo
    successCriteria:
    - condition: $statusCode == 200
    - condition: $response.body#/status == "success"
    - condition: $response.body#/data/reloadConfigSuccess == true
    outputs:
      reloadConfigSuccess: $response.body#/data/reloadConfigSuccess
      lastConfigTime: $response.body#/data/lastConfigTime
      startTime: $response.body#/data/startTime
  - stepId: captureConfigAfter
    description: >-
      Re-read the loaded configuration and confirm the running server is serving
      the new content rather than the before image.
    operationId: getConfig
    successCriteria:
    - condition: $statusCode == 200
    - condition: $response.body#/status == "success"
    outputs:
      configYamlAfter: $response.body#/data/yaml
  - stepId: confirmRulesLoaded
    description: >-
      List the rule groups now loaded. A reload can succeed on the main
      configuration while a referenced rule file quietly contributes nothing, so
      confirming the expected groups are present closes the loop.
    operationId: getRules
    parameters:
    - name: type
      in: query
      value: $inputs.ruleType
    successCriteria:
    - condition: $statusCode == 200
    - condition: $response.body#/status == "success"
    outputs:
      ruleGroups: $response.body#/data/groups
      firstGroupName: $response.body#/data/groups/0/name
      firstGroupFile: $response.body#/data/groups/0/file
      firstGroupLastEvaluation: $response.body#/data/groups/0/lastEvaluation
  outputs:
    configYamlBefore: $steps.captureConfigBefore.outputs.configYamlBefore
    configYamlAfter: $steps.captureConfigAfter.outputs.configYamlAfter
    reloadConfigSuccess: $steps.confirmReloadSucceeded.outputs.reloadConfigSuccess
    lastConfigTime: $steps.confirmReloadSucceeded.outputs.lastConfigTime
    ruleGroups: $steps.confirmRulesLoaded.outputs.ruleGroups