Version: v1.8

Definition

OAM Definitions are basic building blocks of the KubeVela platform. A definition encapsulates an arbitrarily complex automation as a lego style module, then safely shared, discovered and be used to compose an Application delivered by any KubeVela engine.

With the help of OAM definitions, the platform builder can easily adopt infrastructure capabilities from ecosystem, hide complexity with their best practices without losing any extensibility, the most important is the application won’t be affected even if we change the capability providers.

alt

The picture above is an example of helm definition, we can use FluxCD or ArgoCD as Helm capability provider. Once the platform builder registered the definition written by CUE, the end user can discover the helm capability and define application by referring to the type.

There’re mainly four types of definitions in KubeVela, they’re ComponentDefinition, TraitDefinition, PolicyDefinition and WorkflowStepDefinition, corresponding to the application concepts. As an end user, you can get out-of-box definitions from KubeVela community.

There’re two sources to get out-of-box definitions:

A definition’s lifecycle usually has 3 stages:

When definitions installed in the system, they can be discovered by end user immediately.

  • Check the definition list
  1. vela def list

expected output

  1. NAME TYPE NAMESPACE DESCRIPTION
  2. webservice ComponentDefinition vela-system Describes long-running, scalable, containerized services
  3. that have a stable network endpoint to receive external
  4. network traffic from customers.
  5. gateway TraitDefinition vela-system Enable public web traffic for the component, the ingress API
  6. matches K8s v1.20+.
  7. health PolicyDefinition vela-system Apply periodical health checking to the application.
  8. notification WorkflowStepDefinition vela-system Send message to webhook
  9. ...snip...
  • Show property details of one definition
  1. vela show webservice

expected output

  1. # Properties
  2. +------------------+-------------------------------------------------------------------------------------------+-----------------------------------+----------+---------+
  3. | NAME | DESCRIPTION | TYPE | REQUIRED | DEFAULT |
  4. +------------------+-------------------------------------------------------------------------------------------+-----------------------------------+----------+---------+
  5. | cmd | Commands to run in the container | []string | false | |
  6. | env | Define arguments by using environment variables | [[]env](#env) | false | |
  7. | labels | Specify the labels in the workload | map[string]string | false | |
  8. | annotations | Specify the annotations in the workload | map[string]string | false | |
  9. | image | Which image would you like to use for your service | string | true | |
  10. | ports | Which ports do you want customer traffic sent to, defaults to 80 | [[]ports](#ports) | false | |
  11. +------------------+-------------------------------------------------------------------------------------------+-----------------------------------+----------+---------+
  12. ...snip...

You can also view the details from documentation website, the following command will launch a server and invoke your browser automatically:

  1. vela show webservice --web
  • Discover in UI console (velaux)

Once installed, these definitions can be discovered and generated with forms by the UI console automatically. More than that, you can even customize the layout by defining the ui schema like below.

alt

Using definition on UI console is very straight forward, just click along with the application creation process.

  1. Create Application and choose Component type which is actually choosing which component definition to use.
  2. Fill the properties of component is actually fill the parameter of component definition.
  3. The same step for trait, policy and workflow.

Finally, the UI console will compose the whole deployment plan in the format of OAM like below, then KubeVela controller will take care of the rest things:

  1. apiVersion: core.oam.dev/v1beta1
  2. kind: Application
  3. metadata:
  4. name: first-vela-app
  5. spec:
  6. components:
  7. - name: express-server
  8. type: webservice
  9. properties:
  10. image: oamdev/hello-world
  11. ports:
  12. - port: 8000
  13. expose: true
  14. traits:
  15. - type: scaler
  16. properties:
  17. replicas: 1
  18. policies:
  19. - name: target-default
  20. type: topology
  21. properties:
  22. clusters: ["local"]
  23. namespace: "default"
  24. - name: target-prod
  25. type: topology
  26. properties:
  27. clusters: ["local"]
  28. namespace: "prod"
  29. - name: deploy-ha
  30. type: override
  31. properties:
  32. components:
  33. - type: webservice
  34. traits:
  35. - type: scaler
  36. properties:
  37. replicas: 2
  38. workflow:
  39. steps:
  40. - name: deploy2default
  41. type: deploy
  42. properties:
  43. policies: ["target-default"]
  44. - name: manual-approval
  45. type: suspend
  46. - name: deploy2prod
  47. type: deploy
  48. properties:
  49. policies: ["target-prod", "deploy-ha"]

Use the definition in vela command line works the same, you can compose the application yaml manually and deploy by vela up command.

  1. vela up -f https://kubevela.net/example/applications/first-app.yaml

Application is also one kind of Kubernetes CRD, you can also use kubectl apply or invoke Kubernetes API to integrate with vela application.

Definition - 图3caution

In most cases, you don’t need to customize any definitions unless you’re going to extend the capability of KubeVela. Before that, you should check the built-in definitions and addons to confirm if they can fit your needs.

A new definition is built in a declarative template in CUE configuration language. If you’re not familiar with CUE, you can refer to CUE Basic for some knowledge.

A definition describes the module’s inputs, outputs, operations, and the wiring between them. Here is an example of a simple component definition:

  1. webserver: {
  2. type: "component"
  3. attributes: {}
  4. }
  5. template: {
  6. parameter: {
  7. name: string
  8. image: string
  9. }
  10. output: {
  11. apiVersion: "apps/v1"
  12. kind: "Deployment"
  13. spec: {
  14. containers: [{
  15. name: parameter.name
  16. image: parameter.image
  17. }]
  18. }
  19. }
  20. }

The type defines what kind of definition it is, the parameter defines the inputs, while the output section defines the outputs. You can refer to detail docs about how to manage definition or learn the definition protocol.

Last updated on May 6, 2023 by Tianxin Dong