Principle:Kubeflow Pipelines Pipeline Submission
| Sources | Domains | Last Updated |
|---|---|---|
| KFP Client | ML_Pipelines, Execution | 2026-02-13 |
Overview
The process of submitting a compiled or in-memory pipeline to a Kubeflow Pipelines backend for execution on a Kubernetes cluster.
Description
Pipeline submission connects the client-side authoring environment to the server-side execution engine. The KFP Client communicates with the KFP API server to create runs, pass runtime arguments, and optionally wait for completion. This bridges development and production by enabling programmatic pipeline execution.
Usage
Use when ready to execute a pipeline on a KFP deployment, either locally via port-forward or remotely via API endpoint.
Theoretical Basis
Client-server pattern for job submission. The client serializes the pipeline, sends it to the API server, which schedules it on the Kubernetes cluster via Argo Workflows.