github.com/pachyderm/pachyderm@v1.13.4/doc/docs/master/how-tos/developer-workflow/ci-cd-integration.md (about)

     1  # CI/CD Integration
     2  
     3  Pachyderm is a powerful system for providing data
     4  provenance and scalable processing to data
     5  scientists and engineers. You can make it even
     6  more powerful by integrating it with your existing
     7  continuous integration and continuous deployment (CI/CD)
     8  workflows and systems. If you are just starting to use Pachyderm
     9  and not setting up automation for your Pachyderm build
    10  processes, see [Working with Pipelines](working-with-pipelines.md).
    11  
    12  The following diagram demonstrates automated Pachyderm
    13  development workflow with CI:
    14  
    15  ![Developer Workflow](../../assets/images/d_developer_workflow102.svg)
    16  
    17  Although initial CI setup might require extra effort on your side,
    18  in the long run, it brings significant benefits to your team,
    19  including the following:
    20  
    21  * Simplified workflow for data scientists. Data scientists do not need to be
    22  aware of the complexity of the underlying containerized infrastructure. They
    23  can follow an established Git process, and the CI platform takes care of the
    24  Docker build and push process behind the scenes.
    25  
    26  * Your CI platform can run additional unit tests against the submitted
    27  code before creating the build.
    28  
    29  * Flexibility in tagging Docker images, such as specifying a custom name
    30  and tag or using the commit SHA for tagging.
    31  
    32  
    33  ## CI Workflow
    34  
    35  The CI workflow includes the following steps:
    36  
    37  1. A new commit triggers a Git hook.
    38  
    39     Typically, Pachyderm users store the following artifacts in a
    40     Git repository:
    41     * A Dockerfile that you use to build local images.
    42     * A `pipeline.json` specification file that you can use in a `Makefile` to create local builds, as well as in the CI/CD workflows.
    43     * The code that performs data transformations.
    44  
    45     A [commit hook in Git](https://git-scm.com/book/en/v2/Customizing-Git-Git-Hooks)
    46     for your repository triggers the CI/CD process. It uses the
    47     information in your pipeline specification for subsequent steps.
    48  
    49  1. Build an image.
    50  
    51     Your CI process automatically starts the build of a Docker container
    52     image based on your code and the Dockerfile.
    53  
    54  1. Push the image tagged with commit ID to an image registry.
    55  
    56     Your CI process pushes a Docker image created in Step 2 to your preferred
    57     image registry. When a data scientist submits their code to Git, a CI
    58     process uses the Dockerfile in the repository to build, tag with a Git
    59     commit SHA, and push the container to your image registry.
    60  
    61  1. Update the pipeline spec with the tagged image.
    62  
    63     In this step, your CI/CD infrastructure uses your updated `pipeline.json`
    64     specification and fills in the Git commit
    65     SHA for the version of the image that must be used in this pipeline.
    66     Then, it runs the `pachctl update pipeline` command to push the
    67     updated pipeline specification to Pachyderm. After that,
    68     Pachyderm pulls a new image from the registry automatically.
    69     When the production pipeline is updated with the `pipeline.json`
    70     file that has the correct image tag in it, Pachyderm restarts all pods
    71     for this pipeline with the new image automatically.
    72  
    73  
    74  ## GitHub Actions
    75  [GitHub actions](github.com/features/actions) are a convenient way to kick off workflows and perform integration. These can be used to:
    76  
    77  * Manually trigger a pipeline build, or
    78  * Automatically build a pipeline from a commit or pull request.
    79  
    80  In our [example](https://github.com/pachyderm/pachyderm/tree/workflows/examples/workflows/github-actions), we show how to use the Pachyderm GitHub Action to incorporate Pachyderm functions to run on a Pull Request or at other points during development.
    81  
    82  
    83  <!-- ## Jenkins Integration -->