|
| 1 | +:_mod-docs-content-type: ASSEMBLY |
| 2 | +include::_attributes/common-attributes.adoc[] |
| 3 | +[id="about-kueue"] |
| 4 | += About {product-title} |
| 5 | +:context: about-kueue |
| 6 | + |
| 7 | +toc::[] |
| 8 | + |
| 9 | +{product-title} is a Kubernetes-native system that manages quotas and how jobs consume them. Kueue decides when a job waits, when a job is admitted to start by creating pods, and when a job should be _preempted_ (active pods should be deleted). |
| 10 | + |
| 11 | +Kueue is compatible with environments that use heterogeneous, elastic resources. This means that the environment has many different resource types, and those resources are capable of dynamic scaling. |
| 12 | + |
| 13 | +{product-title} is based on the link::https://kueue.sigs.k8s.io/docs/[Kueue] open source project. |
| 14 | + |
| 15 | +// Personas |
| 16 | +[id="about-kueue-personas"] |
| 17 | +== {kueue-s} personas |
| 18 | + |
| 19 | +Different personas exist in a {kueue-s} workflow. |
| 20 | + |
| 21 | +Batch administrators:: Batch administrators manage the cluster infrastructure and establish quotas and queues. |
| 22 | +Batch users:: Batch users run workloads on the cluster. Examples of batch users might be researchers, AI/ML engineers, or data scientists. |
| 23 | +Serving users:: Serving users run workloads on the cluster. For example, to expose a trained AI/ML model for inference. |
| 24 | +Platform developers:: Platform developers integrate Kueue with other software. They might also contribute to the Kueue open source project. |
| 25 | + |
| 26 | +// How it works |
| 27 | +[id="about-kueue-workflow"] |
| 28 | +== {kueue-s} workflow |
| 29 | +// TODO: add diagram? |
| 30 | + |
| 31 | +The {kueue-s} workflow can be described at a high level as follows: |
| 32 | + |
| 33 | +. Batch administrators create and configure resource flavors and queues. |
| 34 | +. User personas create jobs on the cluster. |
| 35 | +. The `kube-apiserver` component validates and accepts job data. |
| 36 | +. Kueue admits jobs based on configured options, such as order or quota. |
| 37 | +. The job controller creates pods. |
| 38 | +. The `kube-scheduler` component assigns pods to a node in the cluster. |
| 39 | +. The `cluster-autoscaler` component provisions more nodes as required. |
0 commit comments