Generate CRD Manifest

Property	Type	Description
kind	string	-
clusters	array	-
nameParam	string	-
apiVersion	string	-
ownerParam	any	-
parameters	object	-
excludeParams	array	-
namespaceParam	string	-
removeEmptyParams	boolean	-

Property	Type	Description	Required
manifest	string	-
filePaths	array	-

Generate a basic CRD for a single cluster and publish to GitHub

Creates one custom resource manifest using provided name and namespace, then publishes the workspace to GitHub. Use this for a simple one-cluster deployment alongside a base repo structure set up with fetch:template and published via publish:github.

Copy

steps:
  - id: fetchBase
    action: fetch:template
    input:
      url: ./skeletons/k8s-manifests
      targetPath: .
      values:
        serviceName: ${{ parameters.name }}

  - id: generateCrd
    action: terasky:crd-template
    input:
      ownerParam: ${{ parameters.owner }}
      parameters:
        team: "payments"
        tier: "backend"
        replicas: 2
        image: "ghcr.io/example/widget:1.2.3"
        featureFlags:
          enableAudit: true
          enableCache: false
      nameParam: ${{ parameters.name }}
      namespaceParam: ${{ parameters.namespace }}
      excludeParams:
        - internalNotes
      apiVersion: ${{ parameters.apiVersion }} # e.g. "example.com/v1"
      kind: ${{ parameters.kind }}             # e.g. "Widget"
      clusters:
        - ${{ parameters.cluster }}            # e.g. "dev-us-east-1"

  - id: publish
    action: publish:github
    input:
      repoUrl: ${{ parameters.repoUrl }}       # e.g. "github.com?owner=acme&repo=widget-crds"
      defaultBranch: main
      description: "CRD manifests for ${{ parameters.name }}"

Generate CRDs for multiple clusters and prune empty fields

Creates manifests for several clusters while removing empty values from the parameter map. Use when some inputs are optional and should not appear in the rendered manifest. The repo is prepared with fetch:template and later published with publish:github.

Copy

steps:
  - id: scaffoldRepo
    action: fetch:template
    input:
      url: ./skeletons/manifest-repo
      targetPath: .
      values:
        name: ${{ parameters.name }}

  - id: crdMultiCluster
    action: terasky:crd-template
    input:
      ownerParam: ${{ parameters.owner }}
      parameters:
        replicas: ${{ parameters.replicas }}   # e.g. 1, 2, 3
        config:
          url: ${{ parameters.configUrl }}     # e.g. "https://api.example.com"
          timeoutSeconds: 30
          extra: {}                            # pruned when removeEmptyParams is true
        notes: ""                               # pruned
        tags:
          - "staging"
      nameParam: ${{ parameters.name }}
      namespaceParam: ${{ parameters.namespace }}
      excludeParams:
        - notes
      apiVersion: "apps.example.com/v1alpha1"
      kind: "Widget"
      removeEmptyParams: true
      clusters: ${{ parameters.clusters }}     # e.g. ["dev", "staging", "prod"]

  - id: publishRepo
    action: publish:github
    input:
      repoUrl: ${{ parameters.repoUrl }}
      defaultBranch: main
      description: "Multi-cluster CRD manifests for ${{ parameters.name }}"

Exclude non-spec inputs and publish to GitLab

Generates a manifest while excluding fields not meant to be rendered into the CRD. Publishes to GitLab using publish:gitlab after initializing the workspace with fetch:template.

Copy

steps:
  - id: init
    action: fetch:template
    input:
      url: ./skeletons/k8s-crd-repo
      targetPath: .
      values:
        component: ${{ parameters.name }}

  - id: crdExclude
    action: terasky:crd-template
    input:
      ownerParam: ${{ parameters.owner }}
      parameters:
        replicas: 1
        image: "registry.gitlab.com/acme/${{ parameters.name }}:0.4.0"
        rolloutStrategy: "RollingUpdate"
        debug: ${{ parameters.debug }}         # exclude this below
        documentationUrl: ${{ parameters.docsUrl }}
      nameParam: ${{ parameters.name }}
      namespaceParam: ${{ parameters.namespace }}
      excludeParams:
        - debug
        - documentationUrl
      apiVersion: "platform.acme.io/v1"
      kind: "RuntimeConfig"
      clusters:
        - "dev-eu-west-1"
        - "prod-eu-west-1"

  - id: publishGitlab
    action: publish:gitlab
    input:
      repoUrl: ${{ parameters.repoUrl }}       # e.g. "gitlab.com?owner=platform&repo=runtime-configs"
      defaultBranch: main
      description: "RuntimeConfig CRDs for ${{ parameters.name }}"

Template a ServiceMonitor CRD with realistic spec fields

Creates a ServiceMonitor custom resource for Prometheus Operator across two clusters. This is useful for exposing service metrics with a known CRD type, combined with fetch:template and publish:github.

Copy

steps:
  - id: baseFiles
    action: fetch:template
    input:
      url: ./skeletons/observability-repo
      targetPath: .
      values:
        service: ${{ parameters.name }}

  - id: serviceMonitor
    action: terasky:crd-template
    input:
      ownerParam: ${{ parameters.owner }}
      parameters:
        selector:
          matchLabels:
            app.kubernetes.io/name: ${{ parameters.name }}
        namespaceSelector:
          matchNames:
            - ${{ parameters.namespace }}
        endpoints:
          - port: "http-metrics"
            interval: "30s"
            path: "/metrics"
      nameParam: "${{ parameters.name }}-sm"
      namespaceParam: ${{ parameters.namespace }}
      excludeParams: []
      apiVersion: "monitoring.coreos.com/v1"
      kind: "ServiceMonitor"
      removeEmptyParams: true
      clusters:
        - "dev-us-central1"
        - "prod-us-central1"

  - id: publishObs
    action: publish:github
    input:
      repoUrl: ${{ parameters.repoUrl }}
      defaultBranch: main
      description: "ServiceMonitor resources for ${{ parameters.name }}"

Environment-suffixed names and cluster fan-out

Appends the environment to the resource name and generates one manifest per cluster. Use this pattern when the same CRD should be deployed with environment-qualified names. The workspace is bootstrapped with fetch:template and published via publish:github.

Copy

steps:
  - id: prep
    action: fetch:template
    input:
      url: ./skeletons/crd-repo
      targetPath: .
      values:
        env: ${{ parameters.environment }}

  - id: crdEnvFanOut
    action: terasky:crd-template
    input:
      ownerParam: ${{ parameters.owner }}
      parameters:
        image: "ghcr.io/acme/${{ parameters.name }}:${{ parameters.version }}"
        replicas: ${{ parameters.replicas }}
        resources:
          requests:
            cpu: "100m"
            memory: "128Mi"
          limits:
            cpu: "500m"
            memory: "256Mi"
        debug: false
      nameParam: "${{ parameters.name }}-${{ parameters.environment }}"
      namespaceParam: ${{ parameters.namespace }}
      excludeParams:
        - debug
      apiVersion: "apps.example.com/v1"
      kind: "Widget"
      clusters:
        - "dev-us1"
        - "stg-us1"
        - "prod-us1"

  - id: publishManifests
    action: publish:github
    input:
      repoUrl: ${{ parameters.repoUrl }}
      defaultBranch: main
      description: "Environment-specific CRDs for ${{ parameters.name }}"

Input Schema

Output Schema

Usage Examples

Generate a basic CRD for a single cluster and publish to GitHub

Generate CRDs for multiple clusters and prune empty fields

Exclude non-spec inputs and publish to GitLab

Template a ServiceMonitor CRD with realistic spec fields

Environment-suffixed names and cluster fan-out

Other actions in @terasky/backstage-plugin-scaffolder-backend-module-terasky-utils

Generate Claim Manifest

Catalog Info Cleaner