Kueue is a set of APIs and controller for job queueing. It is a job-level manager that decides when
a job should start (as in pods can be created) and when it should stop (as in active pods should be
deleted). The main design principle for Kueue is to avoid duplicating existing functionality: autoscaling,
pod-to-node scheduling, job lifecycle management and advanced admission control are the responsibility of
core k8s components or commonly accepted frameworks, namely cluster-autoscaler, kube-scheduler and kube-controller-manager
and gatekeeper, respectively.

bit.ly/kueue-apis (please join the mailing list to get access) discusses the
API proposal and a high-level description of how it operates; while bit.ly/kueue-controller-design presents the detailed design of the controller.


Requires k8s 1.22 or newer

You can run Kueue with the following command:

IMG=registry.example.com/kueue:latest make docker-build docker-push deploy

The controller will run in the kueue-system namespace.
Then, you can and apply some of the samples:

kubectl apply -f config/samples/minimal.yaml
kubectl create -f config/samples/sample-job.yaml

Community, discussion, contribution, and support

Learn how to engage with the Kubernetes community on the community page.

You can reach the maintainers of this project at:

Code of conduct

Participation in the Kubernetes community is governed by the Kubernetes Code of Conduct.


View Github