Lightweight Python 3+ client library for Kubernetes (pykube-ng)
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
Henning Jacobs 47f61bbfbd v22.9.0 3 weeks ago
docs -> 2 years ago
pykube v22.9.0 3 weeks ago
tests fix: resolve KeyError on watch failures 3 weeks ago
.flake8 Switch to poetry & black (#51) 3 years ago
.gitignore update dependencies 3 weeks ago
.pre-commit-config.yaml update dependencies 3 months ago
.tools-version #4 ADD certificate authority env variable to override CA used to validate http calls 2 years ago
.travis.yml Switch to poetry & black (#51) 3 years ago changelog for 0.17 and 0.18 4 years ago
LICENSE initial port of code 7 years ago
Makefile add pre-commit hooks (#52) 3 years ago
README.rst -> 2 years ago
poetry.lock update dependencies 3 weeks ago
pyproject.toml v22.9.0 3 weeks ago



Build status

Code Coverage


PyPI version

Python versions

Apache License


Pykube (pykube-ng) is a lightweight Python 3.6+ client library for Kubernetes.

This is a fork of kelproject/pykube which is no longer maintained (archived). Here the original text of the pykube README:

Kel is an open source Platform as a Service (PaaS) from Eldarion, Inc. that makes it easy to manage web application deployment and hosting through the entire lifecycle from development through testing to production. It adds components and tools on top of Kubernetes that help developers manage their application infrastructure. Kel builds on Eldarion's 7+ years experience running one of the leading Python and Django PaaSes. For more information about Kel, see or follow us on Twitter @projectkel.


  • HTTP interface using requests using kubeconfig for authentication
  • Python native querying of Kubernetes API objects


To install pykube, use pip:

pip install pykube-ng

Interactive Console

The pykube library module can be run as an interactive console locally for quick exploration. It will automatically load ~/.kube/config to provide the api object, and it loads pykube classes (Deployment, Pod, ..) into local context:

python3 -m pykube
>>> [ for d in Deployment.objects(api)]


Query for all ready pods in a custom namespace:

import operator
import pykube

api = pykube.HTTPClient(pykube.KubeConfig.from_file())
pods = pykube.Pod.objects(api).filter(namespace="gondor-system")
ready_pods = filter(operator.attrgetter("ready"), pods)

Access any attribute of the Kubernetes object:

pod = pykube.Pod.objects(api).filter(namespace="gondor-system").get(name="my-pod")

Selector query:

pods = pykube.Pod.objects(api).filter(
    selector={"": {"api-web", "api-worker"}},
pending_pods = pykube.objects.Pod.objects(api).filter(
    field_selector={"status.phase": "Pending"}

Watch query:

watch = pykube.Job.objects(api, namespace="gondor-system")
watch = watch.filter(field_selector={"": "my-job"}).watch()

# watch is a generator:
for watch_event in watch:
    print(watch_event.type) # 'ADDED', 'DELETED', 'MODIFIED'
    print(watch_event.object) # pykube.Job object

Create a Deployment:

obj = {
    "apiVersion": "apps/v1",
    "kind": "Deployment",
    "metadata": {
        "name": "my-deploy",
        "namespace": "gondor-system"
    "spec": {
        "replicas": 3,
        "selector": {
            "matchLabels": {
                "app": "nginx"
        "template": {
            "metadata": {
                "labels": {
                    "app": "nginx"
            "spec": {
                "containers": [
                        "name": "nginx",
                        "image": "nginx",
                        "ports": [
                            {"containerPort": 80}
pykube.Deployment(api, obj).create()

Delete a Deployment:

obj = {
    "apiVersion": "apps/v1",
    "kind": "Deployment",
    "metadata": {
        "name": "my-deploy",
        "namespace": "gondor-system"
pykube.Deployment(api, obj).delete()

Check server version:

api = pykube.HTTPClient(pykube.KubeConfig.from_file())


  • Python 3.6+
  • requests (included in install_requires)
  • PyYAML (included in install_requires)

Local Development

You can run pykube against your current kubeconfig context, e.g. local Minikube:

poetry install
poetry run python3
>>> import pykube
>>> config = pykube.KubeConfig.from_file()
>>> api = pykube.HTTPClient(config)
>>> list(pykube.Deployment.objects(api))

To run PEP8 (flake8) checks and unit tests including coverage report:

make test


The code in this project is licensed under the Apache License, version 2.0 (included in this repository under LICENSE).


Easiest way to contribute is to provide feedback! We would love to hear what you like and what you think is missing. Create an issue or ping try_except_ on Twitter.

PRs are welcome. Please also have a look at issues labeled with "help wanted".

Code of Conduct

In order to foster a kind, inclusive, and harassment-free community, this project follows the Contributor Covenant Code of Conduct.