Version v0.5 of the documentation is no longer actively maintained. The site that you are currently viewing is an archived snapshot. For up-to-date documentation, see the latest version.

Overview of Kubeflow Pipelines

Overview of Kubeflow Pipelines

Kubeflow Pipelines is a platform for building and deploying portable, scalable machine learning (ML) workflows based on Docker containers.


Run your first pipeline by following the pipelines quickstart guide.

What is Kubeflow Pipelines?

The Kubeflow Pipelines platform consists of:

  • A user interface (UI) for managing and tracking experiments, jobs, and runs.
  • An engine for scheduling multi-step ML workflows.
  • An SDK for defining and manipulating pipelines and components.
  • Notebooks for interacting with the system using the SDK.

The following are the goals of Kubeflow Pipelines:

  • End-to-end orchestration: enabling and simplifying the orchestration of machine learning pipelines.
  • Easy experimentation: making it easy for you to try numerous ideas and techniques and manage your various trials/experiments.
  • Easy re-use: enabling you to re-use components and pipelines to quickly create end-to-end solutions without having to rebuild each time.

In Kubeflow v0.1.3 and later, Kubeflow Pipelines is one of the Kubeflow core components. It’s automatically deployed during Kubeflow deployment. You can try it currently with a Kubeflow deployment on GKE in Google Cloud Platform (GCP). See the guide to deploying Kubeflow on GCP.

Due to kubeflow/pipelines#345 and kubeflow/pipelines#337, some non-critical pieces of functionality are currently available only on GKE clusters.

What is a pipeline?

A pipeline is a description of an ML workflow, including all of the components in the workflow and how they combine in the form of a graph. (See the screenshot below showing an example of a pipeline graph.) The pipeline includes the definition of the inputs (parameters) required to run the pipeline and the inputs and outputs of each component.

After developing your pipeline, you can upload and share it on the Kubeflow Pipelines UI.

A pipeline component is a self-contained set of user code, packaged as a Docker image, that performs one step in the pipeline. For example, a component can be responsible for data preprocessing, data transformation, model training, and so on.

See the conceptual guides to pipelines and components.

Example of a pipeline

The screenshots and code below show the pipeline, which creates an XGBoost model using structured data in CSV format. You can see the source code and other information about the pipeline on GitHub.

The runtime execution graph of the pipeline

The screenshot below shows the example pipeline’s runtime execution graph in the Kubeflow Pipelines UI:

XGBoost results on the pipelines UI

The Python code that represents the pipeline

Below is an extract from the Python code that defines the pipeline. You can see the full code on GitHub.

  name='XGBoost Trainer',
  description='A trainer that does end-to-end distributed training for XGBoost models.'
def xgb_train_pipeline(
  delete_cluster_op = DeleteClusterOp('delete-cluster', project, region).apply(gcp.use_gcp_secret('user-gcp-sa'))
  with dsl.ExitHandler(exit_op=delete_cluster_op):
    create_cluster_op = CreateClusterOp('create-cluster', project, region, output).apply(gcp.use_gcp_secret('user-gcp-sa'))

    analyze_op = AnalyzeOp('analyze', project, region, create_cluster_op.output, schema,
                           train_data, '%s/{{}}/analysis' % output).apply(gcp.use_gcp_secret('user-gcp-sa'))

    transform_op = TransformOp('transform', project, region, create_cluster_op.output,
                               train_data, eval_data, target, analyze_op.output,
                               '%s/{{}}/transform' % output).apply(gcp.use_gcp_secret('user-gcp-sa'))

    train_op = TrainerOp('train', project, region, create_cluster_op.output, transform_op.outputs['train'],
                         transform_op.outputs['eval'], target, analyze_op.output, workers,
                         rounds, '%s/{{}}/model' % output).apply(gcp.use_gcp_secret('user-gcp-sa'))

    predict_op = PredictOp('predict', project, region, create_cluster_op.output, transform_op.outputs['eval'],
                           train_op.output, target, analyze_op.output, '%s/{{}}/predict' % output).apply(gcp.use_gcp_secret('user-gcp-sa'))

    confusion_matrix_op = ConfusionMatrixOp('confusion-matrix', predict_op.output,
                                            '%s/{{}}/confusionmatrix' % output).apply(gcp.use_gcp_secret('user-gcp-sa'))

    roc_op = RocOp('roc', predict_op.output, true_label, '%s/{{}}/roc' % output).apply(gcp.use_gcp_secret('user-gcp-sa'))

Pipeline data on the Kubeflow Pipelines UI

The screenshot below shows the Kubeflow Pipelines UI for kicking off a run of the pipeline. The pipeline definition in your code determines which parameters appear in the UI form. The pipeline definition can also set default values for these parameters. The arrows on the screenshot indicate the parameters that do not have useful default values in this particular example:

Starting the XGBoost run on the pipelines UI

Outputs from the pipeline

The following screenshots show examples of the pipeline output visible on the Kubeflow Pipelines UI.

Prediction results:

Prediction output

Confusion matrix:

Confusion matrix

Receiver operating characteristics (ROC) curve:


Architectural overview

Pipelines architectural diagram

At a high level, the execution of a pipeline proceeds as follows:

  • Python SDK: You create components or specify a pipeline using the Kubeflow Pipelines domain-specific language (DSL).
  • DSL compiler: The DSL compiler transforms your pipeline’s Python code into a static configuration (YAML).
  • Pipeline Service: You call the Pipeline Service to create a pipeline run from the static configuration.
  • Kubernetes resources: The Pipeline Service calls the Kubernetes API server to create the necessary Kubernetes resources (CRDs) to run the pipeline.
  • Orchestration controllers: A set of orchestration controllers execute the containers needed to complete the pipeline execution specified by the Kubernetes resources (CRDs). The containers execute within Kubernetes Pods on virtual machines. An example controller is the [Argo Workflow]( controller, which orchestrates task-driven workflows.
  • Artifact storage: The Pods store two kinds of data:

    • Metadata: Experiments, jobs, runs, etc. Also single scalar metrics, generally aggregated for the purposes of sorting and filtering. Kubeflow Pipelines stores the metadata in a MySQL database.
    • Artifacts: Pipeline packages, views, etc. Also large-scale metrics like time series, usually used for investigating an individual run’s performance and for debugging. Kubeflow Pipelines stores the artifacts in an artifact store like Minio server or Cloud Storage.

    The MySQL database and the Minio server are both backed by the Kubernetes PersistentVolume (PV) subsystem.

  • Persistence agent and ML metadata: The Pipeline Persistence Agent watches the Kubernetes resources created by the Pipeline Service and persists the state of these resources in the ML Metadata Service. The Pipeline Persistence Agent records the set of containers that executed as well as their inputs and outputs. The input/output consists of either container parameters or data artifact URIs.

  • Pipeline web server: The Pipeline web server gathers data from various services to display relevant views: the list of pipelines currently running, the history of pipeline execution, the list of data artifacts, debugging information about individual pipeline runs, execution status about individual pipeline runs.

Next steps