kubeflow.org_v1_generated.asciidoc

API Reference

Packages

• kubeflow.org/v1

kubeflow.org/v1

Package v1 is the v1 version of the API.

Package v1 contains API Schema definitions for the kubeflow.org v1 API group

Resource Types

• JAXJob

• JAXJobList

• MPIJob

• MPIJobList

• PaddleJob

• PaddleJobList

• PyTorchJob

• PyTorchJobList

• TFJob

• TFJobList

• XGBoostJob

• XGBoostJobList

Definitions

CleanPodPolicy (string)

CleanPodPolicy describes how to deal with pods when the job is finished.

Appears In:

• MPIJobSpec

• RunPolicy

ElasticPolicy

Appears In:

• PyTorchJobSpec

Field	Description
minReplicas integer	minReplicas is the lower limit for the number of replicas to which the training job can scale down. It defaults to null.
maxReplicas integer	upper limit for the number of pods that can be set by the autoscaler; cannot be smaller than MinReplicas, defaults to null.
rdzvBackend RDZVBackend
rdzvPort integer
rdzvHost string
rdzvId string
rdzvConf RDZVConf array	RDZVConf contains additional rendezvous configuration (<key1>=<value1>,<key2>=<value2>,…​).
standalone boolean	Start a local standalone rendezvous backend that is represented by a C10d TCP store on port 29400. Useful when launching single-node, multi-worker job. If specified --rdzv_backend, --rdzv_endpoint, --rdzv_id are auto-assigned; any explicitly set values are ignored.
nProcPerNode integer	Number of workers per node; supported values: [auto, cpu, gpu, int]. Deprecated: This API is deprecated in v1.7+ Use .spec.nprocPerNode instead.
maxRestarts integer
metrics MetricSpec array	Metrics contains the specifications which are used to calculate the desired replica count (the maximum replica count across all metrics will be used). The desired replica count is calculated with multiplying the ratio between the target value and the current value by the current number of pods. Ergo, metrics used must decrease as the pod count is increased, and vice-versa. See the individual metric source types for more information about how each type of metric must respond. If not set, the HPA will not be created.

JAXJob

JAXJob Represents a JAXJob resource.

Appears In:

• JAXJobList

Field	Description
apiVersion string	kubeflow.org/v1
kind string	JAXJob
kind string	Kind is a string value representing the REST resource this object represents. Servers may infer this from the endpoint the client submits requests to. Cannot be updated. In CamelCase. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#types-kinds
apiVersion string	APIVersion defines the versioned schema of this representation of an object. Servers should convert recognized schemas to the latest internal value, and may reject unrecognized values. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#resources
metadata ObjectMeta	Refer to Kubernetes API documentation for fields of metadata.
spec JAXJobSpec	Specification of the desired state of the JAXJob.
status JobStatus	Most recently observed status of the JAXJob. Read-only (modified by the system).

JAXJobList

JAXJobList is a list of JAXJobs.

Field	Description
apiVersion string	kubeflow.org/v1
kind string	JAXJobList
kind string	Kind is a string value representing the REST resource this object represents. Servers may infer this from the endpoint the client submits requests to. Cannot be updated. In CamelCase. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#types-kinds
apiVersion string	APIVersion defines the versioned schema of this representation of an object. Servers should convert recognized schemas to the latest internal value, and may reject unrecognized values. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#resources
metadata ListMeta	Refer to Kubernetes API documentation for fields of metadata.
items JAXJob array	List of JAXJobs.

JAXJobSpec

JAXJobSpec is a desired state description of the JAXJob.

Appears In:

• JAXJob

Field	Description
runPolicy RunPolicy	RunPolicy encapsulates various runtime policies of the distributed training job, for example how to clean up resources and how long the job can stay active.
jaxReplicaSpecs object (keys:ReplicaType, values:ReplicaSpec)	A map of JAXReplicaType (type) to ReplicaSpec (value). Specifies the JAX cluster configuration. For example, { "Worker": JAXReplicaSpec, }

JobCondition

JobCondition describes the state of the job at a certain point.

Appears In:

• JobStatus

Field	Description
type JobConditionType	Type of job condition.
status ConditionStatus	Status of the condition, one of True, False, Unknown.
reason string	The reason for the condition’s last transition.
message string	A human readable message indicating details about the transition.
lastUpdateTime Time	The last time this condition was updated.
lastTransitionTime Time	Last time the condition transitioned from one status to another.

JobConditionType (string)

JobConditionType defines all kinds of types of JobStatus.

Appears In:

• JobCondition

JobStatus

JobStatus represents the current observed state of the training Job.

Appears In:

• JAXJob

• MPIJob

• PaddleJob

• PyTorchJob

• TFJob

• XGBoostJob

Field	Description
conditions JobCondition array	Conditions is an array of current observed job conditions.
replicaStatuses object (keys:ReplicaType, values:ReplicaStatus)	ReplicaStatuses is map of ReplicaType and ReplicaStatus, specifies the status of each replica.
startTime Time	Represents time when the job was acknowledged by the job controller. It is not guaranteed to be set in happens-before order across separate operations. It is represented in RFC3339 form and is in UTC.
completionTime Time	Represents time when the job was completed. It is not guaranteed to be set in happens-before order across separate operations. It is represented in RFC3339 form and is in UTC.
lastReconcileTime Time	Represents last time when the job was reconciled. It is not guaranteed to be set in happens-before order across separate operations. It is represented in RFC3339 form and is in UTC.

MPIJob

Appears In:

• MPIJobList

Field	Description
apiVersion string	kubeflow.org/v1
kind string	MPIJob
kind string	Kind is a string value representing the REST resource this object represents. Servers may infer this from the endpoint the client submits requests to. Cannot be updated. In CamelCase. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#types-kinds
apiVersion string	APIVersion defines the versioned schema of this representation of an object. Servers should convert recognized schemas to the latest internal value, and may reject unrecognized values. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#resources
metadata ObjectMeta	Refer to Kubernetes API documentation for fields of metadata.
spec MPIJobSpec
status JobStatus

MPIJobList

Field	Description
apiVersion string	kubeflow.org/v1
kind string	MPIJobList
kind string	Kind is a string value representing the REST resource this object represents. Servers may infer this from the endpoint the client submits requests to. Cannot be updated. In CamelCase. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#types-kinds
apiVersion string	APIVersion defines the versioned schema of this representation of an object. Servers should convert recognized schemas to the latest internal value, and may reject unrecognized values. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#resources
metadata ListMeta	Refer to Kubernetes API documentation for fields of metadata.
items MPIJob array

MPIJobSpec

Appears In:

• MPIJob

Field	Description
slotsPerWorker integer	Specifies the number of slots per worker used in hostfile. Defaults to 1.
cleanPodPolicy CleanPodPolicy	CleanPodPolicy defines the policy that whether to kill pods after the job completes. Defaults to None.
mpiReplicaSpecs object (keys:ReplicaType, values:ReplicaSpec)	MPIReplicaSpecs contains maps from MPIReplicaType to ReplicaSpec that specify the MPI replicas to run.
mainContainer string	MainContainer specifies name of the main container which executes the MPI code.
runPolicy RunPolicy	RunPolicy encapsulates various runtime policies of the distributed training job, for example how to clean up resources and how long the job can stay active.

PaddleElasticPolicy

Appears In:

• PaddleJobSpec

Field	Description
minReplicas integer	minReplicas is the lower limit for the number of replicas to which the training job can scale down. It defaults to null.
maxReplicas integer	upper limit for the number of pods that can be set by the autoscaler; cannot be smaller than MinReplicas, defaults to null.
maxRestarts integer	MaxRestarts is the limit for restart times of pods in elastic mode.
metrics MetricSpec array	Metrics contains the specifications which are used to calculate the desired replica count (the maximum replica count across all metrics will be used). The desired replica count is calculated with multiplying the ratio between the target value and the current value by the current number of pods. Ergo, metrics used must decrease as the pod count is increased, and vice-versa. See the individual metric source types for more information about how each type of metric must respond. If not set, the HPA will not be created.

PaddleJob

PaddleJob Represents a PaddleJob resource.

Appears In:

• PaddleJobList

Field	Description
apiVersion string	kubeflow.org/v1
kind string	PaddleJob
kind string	Kind is a string value representing the REST resource this object represents. Servers may infer this from the endpoint the client submits requests to. Cannot be updated. In CamelCase. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#types-kinds
apiVersion string	APIVersion defines the versioned schema of this representation of an object. Servers should convert recognized schemas to the latest internal value, and may reject unrecognized values. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#resources
metadata ObjectMeta	Refer to Kubernetes API documentation for fields of metadata.
spec PaddleJobSpec	Specification of the desired state of the PaddleJob.
status JobStatus	Most recently observed status of the PaddleJob. Read-only (modified by the system).

PaddleJobList

PaddleJobList is a list of PaddleJobs.

Field	Description
apiVersion string	kubeflow.org/v1
kind string	PaddleJobList
kind string	Kind is a string value representing the REST resource this object represents. Servers may infer this from the endpoint the client submits requests to. Cannot be updated. In CamelCase. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#types-kinds
apiVersion string	APIVersion defines the versioned schema of this representation of an object. Servers should convert recognized schemas to the latest internal value, and may reject unrecognized values. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#resources
metadata ListMeta	Refer to Kubernetes API documentation for fields of metadata.
items PaddleJob array	List of PaddleJobs.

PaddleJobSpec

PaddleJobSpec is a desired state description of the PaddleJob.

Appears In:

• PaddleJob

Field	Description
runPolicy RunPolicy	RunPolicy encapsulates various runtime policies of the distributed training job, for example how to clean up resources and how long the job can stay active.
elasticPolicy PaddleElasticPolicy	ElasticPolicy holds the elastic policy for paddle job.
paddleReplicaSpecs object (keys:ReplicaType, values:ReplicaSpec)	A map of PaddleReplicaType (type) to ReplicaSpec (value). Specifies the Paddle cluster configuration. For example, { "Master": PaddleReplicaSpec, "Worker": PaddleReplicaSpec, }

PyTorchJob

PyTorchJob Represents a PyTorchJob resource.

Appears In:

• PyTorchJobList

Field	Description
apiVersion string	kubeflow.org/v1
kind string	PyTorchJob
kind string	Kind is a string value representing the REST resource this object represents. Servers may infer this from the endpoint the client submits requests to. Cannot be updated. In CamelCase. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#types-kinds
apiVersion string	APIVersion defines the versioned schema of this representation of an object. Servers should convert recognized schemas to the latest internal value, and may reject unrecognized values. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#resources
metadata ObjectMeta	Refer to Kubernetes API documentation for fields of metadata.
spec PyTorchJobSpec	Specification of the desired state of the PyTorchJob.
status JobStatus	Most recently observed status of the PyTorchJob. Read-only (modified by the system).

PyTorchJobList

PyTorchJobList is a list of PyTorchJobs.

Field	Description
apiVersion string	kubeflow.org/v1
kind string	PyTorchJobList
kind string	Kind is a string value representing the REST resource this object represents. Servers may infer this from the endpoint the client submits requests to. Cannot be updated. In CamelCase. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#types-kinds
apiVersion string	APIVersion defines the versioned schema of this representation of an object. Servers should convert recognized schemas to the latest internal value, and may reject unrecognized values. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#resources
metadata ListMeta	Refer to Kubernetes API documentation for fields of metadata.
items PyTorchJob array	List of PyTorchJobs.

PyTorchJobSpec

PyTorchJobSpec is a desired state description of the PyTorchJob.

Appears In:

• PyTorchJob

Field	Description
runPolicy RunPolicy	RunPolicy encapsulates various runtime policies of the distributed training job, for example how to clean up resources and how long the job can stay active.
elasticPolicy ElasticPolicy
pytorchReplicaSpecs object (keys:ReplicaType, values:ReplicaSpec)	A map of PyTorchReplicaType (type) to ReplicaSpec (value). Specifies the PyTorch cluster configuration. For example, { "Master": PyTorchReplicaSpec, "Worker": PyTorchReplicaSpec, }
nprocPerNode string	Number of workers per node; supported values: [auto, cpu, gpu, int]. For more, https://github.com/pytorch/pytorch/blob/26f7f470df64d90e092081e39507e4ac751f55d6/torch/distributed/run.py#L629-L658. Defaults to auto.

RDZVBackend (string)

Appears In:

• ElasticPolicy

RDZVConf

Appears In:

• ElasticPolicy

Field	Description
key string
value string

ReplicaSpec

ReplicaSpec is a description of the replica

Appears In:

• JAXJobSpec

• MPIJobSpec

• PaddleJobSpec

• PyTorchJobSpec

• TFJobSpec

• XGBoostJobSpec

Field	Description
replicas integer	Replicas is the desired number of replicas of the given template. If unspecified, defaults to 1.
template PodTemplateSpec	Template is the object that describes the pod that will be created for this replica. RestartPolicy in PodTemplateSpec will be overide by RestartPolicy in ReplicaSpec
restartPolicy RestartPolicy	Restart policy for all replicas within the job. One of Always, OnFailure, Never and ExitCode. Default to Never.

ReplicaStatus

ReplicaStatus represents the current observed state of the replica.

Appears In:

• JobStatus

Field	Description
active integer	The number of actively running pods.
succeeded integer	The number of pods which reached phase Succeeded.
failed integer	The number of pods which reached phase Failed.
labelSelector LabelSelector	Deprecated: Use Selector instead
selector string	A Selector is a label query over a set of resources. The result of matchLabels and matchExpressions are ANDed. An empty Selector matches all objects. A null Selector matches no objects.

ReplicaType (string)

ReplicaType represents the type of the replica. Each operator needs to define its own set of ReplicaTypes.

Appears In:

• JAXJobSpec

• JobStatus

• MPIJobSpec

• PaddleJobSpec

• PyTorchJobSpec

• TFJobSpec

• XGBoostJobSpec

RestartPolicy (string)

RestartPolicy describes how the replicas should be restarted. Only one of the following restart policies may be specified. If none of the following policies is specified, the default one is RestartPolicyAlways.

Appears In:

• ReplicaSpec

RunPolicy

RunPolicy encapsulates various runtime policies of the distributed training job, for example how to clean up resources and how long the job can stay active.

Appears In:

• JAXJobSpec

• MPIJobSpec

• PaddleJobSpec

• PyTorchJobSpec

• TFJobSpec

• XGBoostJobSpec

Field	Description
cleanPodPolicy CleanPodPolicy	CleanPodPolicy defines the policy to kill pods after the job completes. Default to None.
ttlSecondsAfterFinished integer	TTLSecondsAfterFinished is the TTL to clean up jobs. It may take extra ReconcilePeriod seconds for the cleanup, since reconcile gets called periodically. Default to infinite.
activeDeadlineSeconds integer	Specifies the duration in seconds relative to the startTime that the job may be active before the system tries to terminate it; value must be positive integer.
backoffLimit integer	Optional number of retries before marking this job failed.
schedulingPolicy SchedulingPolicy	SchedulingPolicy defines the policy related to scheduling, e.g. gang-scheduling
suspend boolean	suspend specifies whether the Job controller should create Pods or not. If a Job is created with suspend set to true, no Pods are created by the Job controller. If a Job is suspended after creation (i.e. the flag goes from false to true), the Job controller will delete all active Pods and PodGroups associated with this Job. Users must design their workload to gracefully handle this. Suspending a Job will reset the StartTime field of the Job. Defaults to false.

SchedulingPolicy

SchedulingPolicy encapsulates various scheduling policies of the distributed training job, for example minAvailable for gang-scheduling.

Appears In:

• RunPolicy

Field	Description
minAvailable integer
queue string
minResources Quantity
priorityClass string
scheduleTimeoutSeconds integer

SuccessPolicy (string)

SuccessPolicy is the success policy.

Appears In:

• TFJobSpec

TFJob

TFJob represents a TFJob resource.

Appears In:

• TFJobList

Field	Description
apiVersion string	kubeflow.org/v1
kind string	TFJob
kind string	Kind is a string value representing the REST resource this object represents. Servers may infer this from the endpoint the client submits requests to. Cannot be updated. In CamelCase. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#types-kinds
apiVersion string	APIVersion defines the versioned schema of this representation of an object. Servers should convert recognized schemas to the latest internal value, and may reject unrecognized values. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#resources
metadata ObjectMeta	Refer to Kubernetes API documentation for fields of metadata.
spec TFJobSpec	Specification of the desired state of the TFJob.
status JobStatus	Most recently observed status of the TFJob. Populated by the system. Read-only.

TFJobList

TFJobList is a list of TFJobs.

Field	Description
apiVersion string	kubeflow.org/v1
kind string	TFJobList
kind string	Kind is a string value representing the REST resource this object represents. Servers may infer this from the endpoint the client submits requests to. Cannot be updated. In CamelCase. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#types-kinds
apiVersion string	APIVersion defines the versioned schema of this representation of an object. Servers should convert recognized schemas to the latest internal value, and may reject unrecognized values. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#resources
metadata ListMeta	Refer to Kubernetes API documentation for fields of metadata.
items TFJob array	List of TFJobs.

TFJobSpec

TFJobSpec is a desired state description of the TFJob.

Appears In:

• TFJob

Field	Description
runPolicy RunPolicy	RunPolicy encapsulates various runtime policies of the distributed training job, for example how to clean up resources and how long the job can stay active.
successPolicy SuccessPolicy	SuccessPolicy defines the policy to mark the TFJob as succeeded. Default to "", using the default rules.
tfReplicaSpecs object (keys:ReplicaType, values:ReplicaSpec)	A map of TFReplicaType (type) to ReplicaSpec (value). Specifies the TF cluster configuration. For example, { "PS": ReplicaSpec, "Worker": ReplicaSpec, }
enableDynamicWorker boolean	A switch to enable dynamic worker

XGBoostJob

XGBoostJob is the Schema for the xgboostjobs API

Appears In:

• XGBoostJobList

Field	Description
apiVersion string	kubeflow.org/v1
kind string	XGBoostJob
kind string	Kind is a string value representing the REST resource this object represents. Servers may infer this from the endpoint the client submits requests to. Cannot be updated. In CamelCase. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#types-kinds
apiVersion string	APIVersion defines the versioned schema of this representation of an object. Servers should convert recognized schemas to the latest internal value, and may reject unrecognized values. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#resources
metadata ObjectMeta	Refer to Kubernetes API documentation for fields of metadata.
spec XGBoostJobSpec
status JobStatus

XGBoostJobList

XGBoostJobList contains a list of XGBoostJob

Field	Description
apiVersion string	kubeflow.org/v1
kind string	XGBoostJobList
kind string	Kind is a string value representing the REST resource this object represents. Servers may infer this from the endpoint the client submits requests to. Cannot be updated. In CamelCase. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#types-kinds
apiVersion string	APIVersion defines the versioned schema of this representation of an object. Servers should convert recognized schemas to the latest internal value, and may reject unrecognized values. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#resources
metadata ListMeta	Refer to Kubernetes API documentation for fields of metadata.
items XGBoostJob array

XGBoostJobSpec

XGBoostJobSpec defines the desired state of XGBoostJob

Appears In:

• XGBoostJob

Field	Description
runPolicy RunPolicy	INSERT ADDITIONAL SPEC FIELDS - desired state of cluster Important: Run "make" to regenerate code after modifying this file
xgbReplicaSpecs object (keys:ReplicaType, values:ReplicaSpec)