apache-airflow-providers-cncf-kubernetes

What it is and what it does

This is an Apache Airflow provider package that bridges Airflow's task orchestration engine with Kubernetes clusters. It allows you to define and execute containerized workloads as Airflow tasks, leveraging Kubernetes' native scheduling, resource management, and pod lifecycle features. The package provides operators and hooks that translate Airflow DAG definitions into Kubernetes pod specifications, handling authentication, namespace management, and async execution through its kubernetes and kubernetes_asyncio dependencies.

The provider is production-stable and actively maintained as part of the official Apache Airflow ecosystem. It requires Airflow 2.11.0 or later and supports Python 3.10 through 3.14. Installation is straightforward via pip on top of an existing Airflow setup, with no compiled dependencies. Use it when you need to run containerized jobs within Kubernetes clusters as part of larger Airflow workflows, or when you want Kubernetes' resource isolation and scheduling alongside Airflow's DAG orchestration.

Use it for:

• Run containerized data processing jobs (ML training, ETL pipelines) on Kubernetes from Airflow DAGs.
• Orchestrate multi-container applications where each task runs as a Kubernetes pod with defined resource limits.
• Integrate Kubernetes-native workloads into Airflow workflows without maintaining separate orchestration systems.
• Execute tasks with dynamic resource allocation by leveraging Kubernetes' autoscaling and node management.
• Deploy Airflow workflows across multiple Kubernetes clusters or namespaces for multi-tenant environments.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Integrates Apache Airflow with Kubernetes, enabling orchestration of containerized workloads and cluster operations through Airflow DAGs.

Yes. This is a production-stable, actively maintained official Airflow provider with no known vulnerabilities, low install friction, and permissive licensing. Install it if you run Airflow and need to orchestrate Kubernetes workloads. Verify your Airflow version meets the >=2.11.0 minimum and that you have Kubernetes cluster access before deploying.

Install

apache-airflow-providers-cncf-kubernetes on PyPI

pip install apache-airflow-providers-cncf-kubernetes

uv add apache-airflow-providers-cncf-kubernetes

poetry add apache-airflow-providers-cncf-kubernetes

Installing apache-airflow-providers-cncf-kubernetes

Before you install

Low install friction; pure Python wheel. Actively maintained with recent release (4 days old). Requires Apache Airflow >=2.11.0 and kubernetes library >=35.0.0,!=36.0.0,<37.0.0; verify your Airflow version meets the minimum.

License in practice

Apache-2.0 permissive license; you may use, modify, and distribute freely with attribution and no warranty.

Quickstart

pip install apache-airflow-providers-cncf-kubernetes

from airflow.providers.cncf.kubernetes.operators.kubernetes_pod import KubernetesPodOperator
from airflow import DAG

with DAG('k8s_example') as dag:
    task = KubernetesPodOperator(
        task_id='run_pod',
        image='my-image:latest',
        namespace='default'
    )

Requires an existing Airflow installation (>=2.11.0) and a Kubernetes cluster or kubeconfig access.

Verify before relying

• Whether async support (kubernetes_asyncio dependency) is automatically enabled or requires explicit configuration.
• Whether the provider handles kubeconfig discovery from standard locations or requires explicit credential setup.

Package facts

Evidence: apache_airflow_providers_cncf_kubernetes-10.21.0-py3-none-any.whl

Keywords: airflow-provider, cncf.kubernetes, airflow, integration

Tags

More Monitoring packages