Plugin Development

KubeOrch supports a plugin system based on Kubernetes Custom Resource Definitions (CRDs). Plugins extend the platform with additional node types, integrations, and capabilities.

Plugin Architecture

Plugins in KubeOrch are registered as CRD extensions. Each plugin:

• Defines a new node type that appears in the workflow canvas
• Specifies the Kubernetes resources it creates
• Provides configuration fields for the settings panel
• Belongs to a category for organization in the plugin marketplace

Plugin Categories

KubeOrch supports 13+ plugin categories:

Category	Description	Examples
virtualization	VM and container runtime	KubeVirt
networking	Network configuration	Cilium, Calico
storage	Storage provisioning	Rook-Ceph, Longhorn
monitoring	Observability and metrics	Prometheus, Grafana
security	Security and policy	Falco, OPA
workflow	Workflow orchestration	Argo Workflows
database	Database operators	CloudNativePG, MongoDB Operator
messaging	Message queues	RabbitMQ, Kafka
backup	Backup and recovery	Velero
ci-cd	CI/CD pipelines	Tekton
ml	Machine learning	Kubeflow
policy	Policy engines	Kyverno
scaling	Auto-scaling	KEDA

Plugin Structure

A plugin is defined as a JSON object with the following structure:

{
  "name": "my-plugin",
  "display_name": "My Custom Plugin",
  "description": "Description of what this plugin does",
  "version": "1.0.0",
  "category": "monitoring",
  "icon": "chart-bar",
  "crd": {
    "group": "example.com",
    "version": "v1",
    "kind": "MyResource",
    "plural": "myresources"
  },
  "config_fields": [
    {
      "name": "replicas",
      "label": "Replicas",
      "type": "number",
      "default": 1,
      "required": true
    },
    {
      "name": "image",
      "label": "Container Image",
      "type": "string",
      "default": "myregistry/myimage:latest",
      "required": true
    },
    {
      "name": "enable_metrics",
      "label": "Enable Metrics",
      "type": "boolean",
      "default": true,
      "required": false
    }
  ],
  "template": "apiVersion: example.com/v1\nkind: MyResource\nmetadata:\n  name: {{.name}}\nspec:\n  replicas: {{.replicas}}\n  image: {{.image}}"
}

Field Types

Plugin configuration fields support these types:

Type	Description	UI Component
string	Text input	Text field
number	Numeric input	Number field
boolean	Toggle	Switch
select	Dropdown selection	Select with options
textarea	Multi-line text	Text area
secret	Sensitive value	Password field

For select fields, add an options array:

{
  "name": "storage_class",
  "label": "Storage Class",
  "type": "select",
  "options": ["standard", "premium-rwo", "premium-rwx"],
  "default": "standard"
}

Creating a Plugin

Step 1: Define the Plugin

Create a plugin definition JSON file:

{
  "name": "redis-cluster",
  "display_name": "Redis Cluster",
  "description": "Deploy a Redis cluster using the Redis Operator",
  "version": "1.0.0",
  "category": "database",
  "icon": "database",
  "crd": {
    "group": "redis.redis.opstreelabs.in",
    "version": "v1beta2",
    "kind": "RedisCluster",
    "plural": "redisclusters"
  },
  "config_fields": [
    {
      "name": "cluster_size",
      "label": "Cluster Size",
      "type": "select",
      "options": ["3", "5", "7"],
      "default": "3",
      "required": true
    },
    {
      "name": "storage_size",
      "label": "Storage Size",
      "type": "string",
      "default": "1Gi",
      "required": true
    },
    {
      "name": "password",
      "label": "Redis Password",
      "type": "secret",
      "required": false
    }
  ],
  "template": "apiVersion: redis.redis.opstreelabs.in/v1beta2\nkind: RedisCluster\nmetadata:\n  name: {{.name}}\nspec:\n  clusterSize: {{.cluster_size}}\n  kubernetesConfig:\n    image: quay.io/opstree/redis:v7.0.12\n  storage:\n    volumeClaimTemplate:\n      spec:\n        accessModes: [\"ReadWriteOnce\"]\n        resources:\n          requests:\n            storage: {{.storage_size}}"
}

Step 2: Register the Plugin

Use the KubeOrch API to register your plugin:

curl -X POST http://localhost:8080/v1/api/plugins \
  -H "Authorization: Bearer <token>" \
  -H "Content-Type: application/json" \
  -d @redis-cluster-plugin.json

Step 3: Enable the Plugin

Users can enable the plugin from the Integrations page or via API:

curl -X POST http://localhost:8080/v1/api/plugins/<plugin-id>/enable \
  -H "Authorization: Bearer <token>"

Step 4: Use in Workflows

Once enabled, the plugin appears as a new node type in the workflow canvas. Users can drag it onto the canvas, configure it in the settings panel, and connect it to other nodes.

Template Syntax

Plugin templates use Go template syntax. Available variables:

Variable	Description
{{.name}}	The node name from the workflow canvas
{{.namespace}}	Target deployment namespace
{{.<field_name>}}	Any config field defined in config_fields

Conditional Logic

spec:
  {{if .enable_metrics}}
  metrics:
    enabled: true
  {{end}}

Default Values

spec:
  replicas: {{or .replicas 1}}

Prerequisites

For a plugin’s CRD to work, the corresponding operator must be installed on the target cluster. Document this in your plugin’s description so users know what to install.

Testing Plugins

1. Register the plugin on a development instance
2. Enable it and add to a workflow
3. Verify the settings panel shows all config fields correctly
4. Deploy to a test cluster and verify the generated Kubernetes resource
5. Check that the resource appears in KubeOrch’s resource monitoring

Best Practices

• Version your plugins — increment the version when changing config fields or templates
• Validate inputs — use required: true for essential fields and provide sensible defaults
• Document prerequisites — list any operators or CRDs that must be installed
• Use meaningful categories — helps users discover your plugin in the marketplace
• Keep templates minimal — only include fields that users need to configure; let the operator handle defaults