Choosing your Runtime Image

The default runtime image used to execute a Bash step is a NodeJS image that matches the OS of your node pool. For a PowerShell step, the default is a .NET Core image. In both cases, the version selected is configurable as part of the Pipelines installation and an administrator may have configured a different default image in your installation. The default is usually most appropriate when your steps will only execute shell commands that are not specific to a language.

For native steps that perform language-specific operations (for example, MvnBuild or GoBuild), the default runtime image will be one appropriate to that language. Similarly, the runtime image chosen will be one that matches the OS of your node pool. Native steps without any language-specific operations will use the same default image as Bash on Ubuntu or CentOS and the same image as PowerShell on Windows Server.

JFrog Pipelines supports Java, C++, Node.js, .NET,, and Go programming languages out of the box. JFrog provides default images in a standard runtime images library, which already have the latest language versions pre-installed. You can use Pipelines to build with other languages by providing a custom image, or installing the language version you need as part of your onStart configuration.

In the runtime section of your YAML configuration you can specify a language and version , which Pipelines will use to select the matching language image. This setting can be configured at a pipeline or step level. As an example, pipeline level configuration is shown below:

pipelines:
  - name: pipeline_1
    configuration:
      runtime:
        type: image
        image:
          auto:
            language: node
            version: "18" # Either "version" or "versions" can be used.
            versions:
              - "18"

If you need an unsupported version of a language, you can either install it in the onStart section, or use a custom Docker image.

Example 1: Specify the default image

pipelines:
  - name: myjfrog_ui_docker
    configuration:
      nodePool: my_cypress
      runtime:
        type: image
        image:
          custom:
            registry: docker
            sourceRepository: apps-docker-local
            name: test.mycompany.io/apps-docker-local/ci/builder-node14
            tag: 2022.8.1-BA-8212-remove-testexecutionreportpaths-property-validation

Example 2: Specify an image for a single step

    steps
      - name: new_cypress
        type: Matrix
        stepMode: Bash
        configuration:
          affinityGroup: group4
          runtime:
            type: image
            image:
              custom:
                registry: docker
                sourceRepository: docker-registry
                name: test.mycompany.io/docker-registry/cypress/base
                tag: 14.18.1

In some cases, our standard images library might not satisfy your requirements, or you might have an internal image you want to use for step execution or you may want to use our standard images library with specific settings. Using your own custom image gives you better control over what is installed on the runtime images. It can also speed up step execution since you can pre-install dependencies into the image.

You can configure a custom runtime image by using the configuration shown below. This can be specified at a pipeline or step level. If specified at both levels, the step configuration will take precedence.

pipelines:
  - name: pipe1
    configuration: 
      runtime:       
        type: image
        image:
          custom:
            name: <string>                        # imageRepo/imageName format
            tag: <string>                         # version tag of the initial version of the Docker image
            options: "-e HOME=/root"          # Docker options
            autoPull: <boolean>                   # default true; pulls image before run
            registry: <integration>       # if custom image is private, an integration is required for authentication
            sourceRepository: <path>      # may be required if registry is Artifactory. e.g. docker-local
            region: <string>              # required if registry is AWS. e.g. us-east-1

If your custom image is private, you will also need to create an integration with credentials for your Docker registry, and set additional values in the config:

pipelines:
  - name: pipe1
    configuration: 
      runtime:       
        type: image
        image:
          custom:
            name: docker/jfrog_win
            tag: latest
            options: "-e HOME=/root"
            autoPull: false
            registry: PSSG_DockerRegistry

pipelines:
  - name: pipe2
    configuration: 
      runtime:       
        type: image
        image:
          custom:
            name: ubuntu
            tag: latest
            registry: myArtifactory
            sourceRepository: docker-local

pipelines:
  - name: pipe3
    configuration: 
      runtime:       
        type: image_aws
        image:
          custom:
            name: 541601.dkr.ecr.us-east-1.amazonaws.com/jfbeta
            tag: latest
            autoPull: false
            registry: AWS
            region: us-east-1