Request
Project to create.
namestringRequired<=63 characters
The unique name of the project within the cluster. Must be a valid Kubernetes name (lowercase, alphanumeric, hyphens allowed).
clusterIdstringRequiredformat: "uuid"
The id of the cluster.
resourceslist of objectsRequired
An array of resource configurations for the project. Each entry defines GPU, CPU, and CPU memory assigned to a specific node pool.
descriptionstringOptional
The description of the project.
schedulingRulesobjectOptional
Scheduling rules that define time limits and idle GPU timeouts for workloads in this org-unit. These rules help manage resource utilization by automatically terminating or preempting idle or long-running workloads.
defaultNodePoolslist of strings or nullOptional
An ordered list of node pools that defines the default scheduling preference for workloads in this project. The Scheduler will attempt to schedule workloads on node pools in this order. This setting is enforced if no node pool list is defined in the workload policy.
nodeTypesobjectOptional
Deprecated: use a workload policy node selector to target node types instead. Node type configurations for different workload types (training, workspace). Specifies which node types are allowed for each workload type. Enforced when no node type is configured in the workload policy.
requestedNamespacestringOptional
The custom Kubernetes namespace name for the project. If not specified, a namespace will be auto-generated using the pattern runai-<project-name>.
enforceRunaiSchedulerboolean or nullOptionalDefaults to true
When set to true, enforces that all pods in this project’s namespace are scheduled using the NVIDIA Run:ai Scheduler (runai-scheduler). Supported for cluster version 2.21 and later.
parentIdstring or nullOptional
The ID of the parent department. When specified, the project is created as a child of this department and inherits its quota, the quota of other projects within the department and its scheduling rules.
Response
Created
descriptionstring
The description of the project.
clusterIdstringformat: "uuid"
The id of the cluster.
parentIdstring
department parent uuid
enforceRunaiSchedulerboolean
a flag to enforce that all pods in this namespace will be scheduled with runai-scheduler. Supported for cluster v2.21+
resourceslist of objects
Resources assigned to this Organization per Node Pool
schedulingRulesobjectOptional
Scheduling rules that define time limits and idle GPU timeouts for workloads in this org-unit. These rules help manage resource utilization by automatically terminating or preempting idle or long-running workloads.
defaultNodePoolslist of strings or nullOptional
An ordered list of node pools that defines the default scheduling preference for workloads in this project. The Scheduler will attempt to schedule workloads on node pools in this order. This setting is enforced if no node pool list is defined in the workload policy.
nodeTypesobjectOptional
Deprecated: use a workload policy node selector to target node types instead. Node type configurations for different workload types (training, workspace). Specifies which node types are allowed for each workload type. Enforced when no node type is configured in the workload policy.
requestedNamespacestring or nullOptional
the project's namespace only if requested
parentobjectOptional
Department or project parent
overtimeDataobjectOptional
