rayapi

Documentation

Index

• type AutoscalerOptions
• type ClusterState
• type HeadGroupSpec
• type HeadInfo
• type JobDeploymentStatus
• type JobStatus
• type JobSubmissionMode
• type RayClusterSpec
• type RayClusterStatus
• type RayJobSpec
• type RayJobStatus
• type ScaleStrategy
• type UpscalingMode
• type WorkerGroupSpec

Types

type AutoscalerOptions

type AutoscalerOptions struct {
	// Resources specifies optional resource request and limit overrides for the autoscaler container.
	// Default values: 500m CPU request and limit. 512Mi memory request and limit.
	Resources *corev1.ResourceRequirements `json:"resources,omitempty"`
	// Image optionally overrides the autoscaler's container image. This override is for provided for autoscaler testing and development.
	Image *string `json:"image,omitempty"`
	// ImagePullPolicy optionally overrides the autoscaler container's image pull policy. This override is for provided for autoscaler testing and development.
	ImagePullPolicy *corev1.PullPolicy `json:"imagePullPolicy,omitempty"`
	// Optional list of environment variables to set in the autoscaler container.
	Env []corev1.EnvVar `json:"env,omitempty"`
	// Optional list of sources to populate environment variables in the autoscaler container.
	EnvFrom []corev1.EnvFromSource `json:"envFrom,omitempty"`
	// Optional list of volumeMounts.  This is needed for enabling TLS for the autoscaler container.
	VolumeMounts []corev1.VolumeMount `json:"volumeMounts,omitempty"`
	// SecurityContext defines the security options the container should be run with.
	// If set, the fields of SecurityContext override the equivalent fields of PodSecurityContext.
	// More info: https://kubernetes.io/docs/tasks/configure-pod-container/security-context/
	SecurityContext *corev1.SecurityContext `json:"securityContext,omitempty"`
	// IdleTimeoutSeconds is the number of seconds to wait before scaling down a worker pod which is not using Ray resources.
	// Defaults to 60 (one minute).
	IdleTimeoutSeconds *int32 `json:"idleTimeoutSeconds,omitempty"`
	// UpscalingMode is "Conservative", "Default", or "Aggressive."
	// Conservative: Upscaling is rate-limited; the number of pending worker pods is at most the size of the Ray cluster.
	// Default: Upscaling is not rate-limited.
	// Aggressive: An alias for Default; upscaling is not rate-limited.
	UpscalingMode *UpscalingMode `json:"upscalingMode,omitempty"`
}

AutoscalerOptions specifies optional configuration for the Ray autoscaler.

type ClusterState

type ClusterState string

The overall state of the Ray cluster.

const (
	Ready     ClusterState = "ready"
	Unhealthy ClusterState = "unhealthy"
	Failed    ClusterState = "failed"
)

type HeadGroupSpec

type HeadGroupSpec struct {
	// ServiceType is Kubernetes service type of the head service. it will be used by the workers to connect to the head pod
	ServiceType corev1.ServiceType `json:"serviceType,omitempty"`
	// HeadService is the Kubernetes service of the head pod.
	HeadService *corev1.Service `json:"headService,omitempty"`
	// EnableIngress indicates whether operator should create ingress object for head service or not.
	EnableIngress *bool `json:"enableIngress,omitempty"`
	// RayStartParams are the params of the start command: node-manager-port, object-store-memory, ...
	RayStartParams map[string]string `json:"rayStartParams"`
	// Template is the exact pod template used in K8s depoyments, statefulsets, etc.
	Template corev1.PodTemplateSpec `json:"template,omitempty"`
}

HeadGroupSpec are the spec for the head pod

type HeadInfo

type HeadInfo struct {
	PodIP     string `json:"podIP,omitempty"`
	ServiceIP string `json:"serviceIP,omitempty"`
}

HeadInfo gives info about head

type JobDeploymentStatus

type JobDeploymentStatus string

JobDeploymentStatus indicates RayJob status including RayCluster lifecycle management and Job submission

const (
	JobDeploymentStatusInitializing                  JobDeploymentStatus = "Initializing"
	JobDeploymentStatusFailedToGetOrCreateRayCluster JobDeploymentStatus = "FailedToGetOrCreateRayCluster"
	JobDeploymentStatusWaitForDashboard              JobDeploymentStatus = "WaitForDashboard"
	JobDeploymentStatusWaitForK8sJob                 JobDeploymentStatus = "WaitForK8sJob"
	JobDeploymentStatusFailedJobDeploy               JobDeploymentStatus = "FailedJobDeploy"
	JobDeploymentStatusRunning                       JobDeploymentStatus = "Running"
	JobDeploymentStatusFailedToGetJobStatus          JobDeploymentStatus = "FailedToGetJobStatus"
	JobDeploymentStatusComplete                      JobDeploymentStatus = "Complete"
	JobDeploymentStatusSuspended                     JobDeploymentStatus = "Suspended"
)

type JobStatus

type JobStatus string

const (
	JobStatusPending   JobStatus = "PENDING"
	JobStatusRunning   JobStatus = "RUNNING"
	JobStatusStopped   JobStatus = "STOPPED"
	JobStatusSucceeded JobStatus = "SUCCEEDED"
	JobStatusFailed    JobStatus = "FAILED"
)

https://docs.ray.io/en/latest/cluster/running-applications/job-submission/jobs-package-ref.html#jobstatus

type JobSubmissionMode

type JobSubmissionMode string

const (
	K8sJobMode JobSubmissionMode = "K8sJobMode" // Submit job via Kubernetes Job
	HTTPMode   JobSubmissionMode = "HTTPMode"   // Submit job via HTTP request
)

type RayClusterSpec

type RayClusterSpec struct {
	// INSERT ADDITIONAL SPEC FIELDS - desired state of cluster
	// Important: Run "make" to regenerate code after modifying this file
	// HeadGroupSpecs are the spec for the head pod
	HeadGroupSpec HeadGroupSpec `json:"headGroupSpec"`
	// WorkerGroupSpecs are the specs for the worker pods
	WorkerGroupSpecs []WorkerGroupSpec `json:"workerGroupSpecs,omitempty"`
	// RayVersion is used to determine the command for the Kubernetes Job managed by RayJob
	RayVersion string `json:"rayVersion,omitempty"`
	// EnableInTreeAutoscaling indicates whether operator should create in tree autoscaling configs
	EnableInTreeAutoscaling *bool `json:"enableInTreeAutoscaling,omitempty"`
	// AutoscalerOptions specifies optional configuration for the Ray autoscaler.
	AutoscalerOptions      *AutoscalerOptions `json:"autoscalerOptions,omitempty"`
	HeadServiceAnnotations map[string]string  `json:"headServiceAnnotations,omitempty"`
	// Suspend indicates whether a RayCluster should be suspended.
	// A suspended RayCluster will have head pods and worker pods deleted.
	Suspend *bool `json:"suspend,omitempty"`
}

RayClusterSpec defines the desired state of RayCluster

type RayClusterStatus

type RayClusterStatus struct {
	// INSERT ADDITIONAL STATUS FIELD - define observed state of cluster
	// Important: Run "make" to regenerate code after modifying this file
	// Status reflects the status of the cluster
	State ClusterState `json:"state,omitempty"`
	// AvailableWorkerReplicas indicates how many replicas are available in the cluster
	AvailableWorkerReplicas int32 `json:"availableWorkerReplicas,omitempty"`
	// DesiredWorkerReplicas indicates overall desired replicas claimed by the user at the cluster level.
	DesiredWorkerReplicas int32 `json:"desiredWorkerReplicas,omitempty"`
	// MinWorkerReplicas indicates sum of minimum replicas of each node group.
	MinWorkerReplicas int32 `json:"minWorkerReplicas,omitempty"`
	// MaxWorkerReplicas indicates sum of maximum replicas of each node group.
	MaxWorkerReplicas int32 `json:"maxWorkerReplicas,omitempty"`
	// LastUpdateTime indicates last update timestamp for this cluster status.
	// +nullable
	LastUpdateTime *metav1.Time `json:"lastUpdateTime,omitempty"`
	// Service Endpoints
	Endpoints map[string]string `json:"endpoints,omitempty"`
	// Head info
	Head HeadInfo `json:"head,omitempty"`
	// Reason provides more information about current State
	Reason string `json:"reason,omitempty"`
	// observedGeneration is the most recent generation observed for this RayCluster. It corresponds to the
	// RayCluster's generation, which is updated on mutation by the API Server.
	// +optional
	ObservedGeneration int64 `json:"observedGeneration,omitempty"`
}

RayClusterStatus defines the observed state of RayCluster

type RayJobSpec

type RayJobSpec struct {
	// Important: Run "make" to regenerate code after modifying this file
	Entrypoint string `json:"entrypoint,omitempty"`
	// Metadata is data to store along with this job.
	Metadata map[string]string `json:"metadata,omitempty"`
	// RuntimeEnvYAML represents the runtime environment configuration
	// provided as a multi-line YAML string.
	RuntimeEnvYAML string `json:"runtimeEnvYAML,omitempty"`
	// If jobId is not set, a new jobId will be auto-generated.
	JobId string `json:"jobId,omitempty"`
	// ShutdownAfterJobFinishes will determine whether to delete the ray cluster once rayJob succeed or failed.
	ShutdownAfterJobFinishes bool `json:"shutdownAfterJobFinishes,omitempty"`
	// TTLSecondsAfterFinished is the TTL to clean up RayCluster.
	// It's only working when ShutdownAfterJobFinishes set to true.
	TTLSecondsAfterFinished *int32 `json:"ttlSecondsAfterFinished,omitempty"`
	// ActiveDeadlineSeconds is the duration in seconds that the RayJob may be active before
	// KubeRay actively tries to terminate the RayJob; value must be positive integer.
	ActiveDeadlineSeconds *int32 `json:"activeDeadlineSeconds,omitempty"`
	// RayClusterSpec is the cluster template to run the job
	RayClusterSpec *RayClusterSpec `json:"rayClusterSpec,omitempty"`
	// clusterSelector is used to select running rayclusters by labels
	ClusterSelector map[string]string `json:"clusterSelector,omitempty"`
	// suspend specifies whether the RayJob controller should create a RayCluster instance
	// If a job is applied with the suspend field set to true,
	// the RayCluster will not be created and will wait for the transition to false.
	// If the RayCluster is already created, it will be deleted.
	// In case of transition to false a new RayCluster will be created.
	Suspend bool `json:"suspend,omitempty"`
	// SubmissionMode specifies how RayJob submits the Ray job to the RayCluster.
	// In "K8sJobMode", the KubeRay operator creates a submitter Kubernetes Job to submit the Ray job.
	// In "HTTPMode", the KubeRay operator sends a request to the RayCluster to create a Ray job.
	// +kubebuilder:default:=K8sJobMode
	SubmissionMode JobSubmissionMode `json:"submissionMode,omitempty"`
	// SubmitterPodTemplate is the template for the pod that will run `ray job submit`.
	SubmitterPodTemplate *corev1.PodTemplateSpec `json:"submitterPodTemplate,omitempty"`
}

type RayJobStatus

type RayJobStatus struct {
	// INSERT ADDITIONAL STATUS FIELD - define observed state of cluster
	// Important: Run "make" to regenerate code after modifying this file
	JobId               string              `json:"jobId,omitempty"`
	RayClusterName      string              `json:"rayClusterName,omitempty"`
	DashboardURL        string              `json:"dashboardURL,omitempty"`
	JobStatus           JobStatus           `json:"jobStatus,omitempty"`
	JobDeploymentStatus JobDeploymentStatus `json:"jobDeploymentStatus,omitempty"`
	Message             string              `json:"message,omitempty"`
	// Represents time when the job was acknowledged by the Ray cluster.
	// It is not guaranteed to be set in happens-before order across separate operations.
	// It is represented in RFC3339 form
	StartTime *metav1.Time `json:"startTime,omitempty"`
	// Represents time when the job was ended.
	EndTime          *metav1.Time     `json:"endTime,omitempty"`
	RayClusterStatus RayClusterStatus `json:"rayClusterStatus,omitempty"`
	// observedGeneration is the most recent generation observed for this RayJob. It corresponds to the
	// RayJob's generation, which is updated on mutation by the API Server.
	// +optional
	ObservedGeneration int64 `json:"observedGeneration,omitempty"`
}

RayJobStatus defines the observed state of RayJob

type ScaleStrategy

type ScaleStrategy struct {
	// WorkersToDelete workers to be deleted
	WorkersToDelete []string `json:"workersToDelete,omitempty"`
}

ScaleStrategy to remove workers

type UpscalingMode

type UpscalingMode string

type WorkerGroupSpec

type WorkerGroupSpec struct {
	// we can have multiple worker groups, we distinguish them by name
	GroupName string `json:"groupName"`
	// Replicas is the number of desired Pods for this worker group. See https://github.com/ray-project/kuberay/pull/1443 for more details about the reason for making this field optional.
	// +kubebuilder:default:=0
	Replicas *int32 `json:"replicas,omitempty"`
	// MinReplicas denotes the minimum number of desired Pods for this worker group.
	// +kubebuilder:default:=0
	MinReplicas *int32 `json:"minReplicas"`
	// MaxReplicas denotes the maximum number of desired Pods for this worker group, and the default value is maxInt32.
	// +kubebuilder:default:=2147483647
	MaxReplicas *int32 `json:"maxReplicas"`
	// NumOfHosts denotes the number of hosts to create per replica. The default value is 1.
	// +kubebuilder:default:=1
	NumOfHosts int32 `json:"numOfHosts,omitempty"`
	// RayStartParams are the params of the start command: address, object-store-memory, ...
	RayStartParams map[string]string `json:"rayStartParams"`
	// Template is a pod template for the worker
	Template corev1.PodTemplateSpec `json:"template"`
	// ScaleStrategy defines which pods to remove
	ScaleStrategy ScaleStrategy `json:"scaleStrategy,omitempty"`
}

WorkerGroupSpec are the specs for the worker pods