feat(autoscaing): support intelligent hpa

config/crd/bases/autoscaling.katalyst.kubewharf.io_virtualworkloads.yaml

@@ -0,0 +1,66 @@
+---
+apiVersion: apiextensions.k8s.io/v1
+kind: CustomResourceDefinition
+metadata:
+  annotations:
+    controller-gen.kubebuilder.io/version: v0.9.0
+  creationTimestamp: null
+  name: virtualworkloads.autoscaling.katalyst.kubewharf.io
+spec:
+  group: autoscaling.katalyst.kubewharf.io
+  names:
+    kind: VirtualWorkload
+    listKind: VirtualWorkloadList
+    plural: virtualworkloads
+    singular: virtualworkload
+  scope: Namespaced
+  versions:
+  - name: v1alpha2
+    schema:
+      openAPIV3Schema:
+        description: VirtualWorkload is the Schema for the virtualworkloads API VirtualWorkload
+          is used to support IHPA's Preview mode, that is, by providing a virtual
+          workload reference so that scaling will not affect the real workload.
+        properties:
+          apiVersion:
+            description: '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'
+            type: string
+          kind:
+            description: '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'
+            type: string
+          metadata:
+            type: object
+          spec:
+            description: VirtualWorkloadSpec defines the desired state of VirtualWorkload
+            properties:
+              replicas:
+                format: int32
+                type: integer
+            required:
+            - replicas
+            type: object
+          status:
+            description: VirtualWorkloadStatus defines the observed state of VirtualWorkload
+            properties:
+              replicas:
+                format: int32
+                type: integer
+              selector:
+                type: string
+            required:
+            - replicas
+            - selector
+            type: object
+        type: object
+    served: true
+    storage: true
+    subresources:
+      scale:
+        labelSelectorPath: .status.selector
+        specReplicasPath: .spec.replicas
+        statusReplicasPath: .status.replicas
+      status: {}

pkg/apis/autoscaling/v1alpha2/ihpa.go

@@ -0,0 +1,220 @@
+package v1alpha2
+
+import (
+	autoscalingv2 "k8s.io/api/autoscaling/v2"
+	corev1 "k8s.io/api/core/v1"
+	"k8s.io/apimachinery/pkg/api/resource"
+	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+
+	"github.com/kubewharf/katalyst-api/pkg/apis/config/v1alpha1"
+)
+
+// +genclient
+// +k8s:deepcopy-gen:interfaces=k8s.io/apimachinery/pkg/runtime.Object
+// +kubebuilder:object:root=true
+// +kubebuilder:subresource:status
+// +kubebuilder:storageversion
+// +kubebuilder:resource:shortName=ihpa
+
+// IntelligentHorizontalPodAutoscaler captures information about a IHPA object
+type IntelligentHorizontalPodAutoscaler struct {
+	metav1.TypeMeta `json:",inline"`
+	// +optional
+	metav1.ObjectMeta `json:"metadata,omitempty"`
+
+	// Spec defines the behavior of a IntelligentHorizontalPodAutoscaler.
+	// +optional
+	Spec IntelligentHorizontalPodAutoscalerSpec `json:"spec,omitempty"`
+
+	// Status represents the current information about a IntelligentHorizontalPodAutoscaler.
+	// +optional
+	Status IntelligentHorizontalPodAutoscalerStatus `json:"status,omitempty"`
+}
+
+// IntelligentHorizontalPodAutoscalerSpec is the specification of the behavior of the autoscaler.
+type IntelligentHorizontalPodAutoscalerSpec struct {
+	// Autoscaler defines the overall scaling configuration.
+	Autoscaler AutoscalerSpec `json:"autoscaler"`
+
+	// ScaleStrategy defines whether to enable IHPA to scale workloads.
+	// IHPA provides a preview mode. In preview mode, IHPA does not modify the number of replicas of the workload.
+	// The default is 'Preview'.
+	// +kubebuilder:default:=Preview
+	ScaleStrategy ScaleStrategyType `json:"scaleStrategy,omitempty"`
+
+	// AlgorithmConfig defines the algorithm configuration. If there is no configuration,
+	// the default configuration will be used.
+	AlgorithmConfig v1alpha1.AlgorithmConfig `json:"algorithmConfig,omitempty"`
+
+	// TimeBounds supports dynamically adjusting HPA Replicas in different time periods,
+	// providing capability similar to CronHPA.
+	TimeBounds []TimeBound `json:"timeBounds,omitempty"`
+}
+
+// IntelligentHorizontalPodAutoscalerStatus describes the runtime state of the autoscaler.
+type IntelligentHorizontalPodAutoscalerStatus struct {
+	// LastHorizontalPodAutoscalerUpdatedTime is the time when the HPA replicas configuration was last refreshed by IHPA.
+	// +optional
+	LastHorizontalPodAutoscalerUpdatedTime *metav1.Time `json:"lastHorizontalPodAutoscalerUpdatedTime,omitempty"`
+
+	// Conditions record the creation or modification time of IHPA's associated resource objects.
+	// +optional
+	// +patchMergeKey=type
+	// +patchStrategy=merge
+	Conditions []IntelligentHorizontalPodAutoscalerCondition `json:"conditions,omitempty" patchStrategy:"merge" patchMergeKey:"type"`
+}
+
+// AutoscalerSpec defines the associated workload, the metrics used, and the scaling behavior.
+type AutoscalerSpec struct {
+	// ScaleTargetRef defines the associated workload.
+	ScaleTargetRef autoscalingv2.CrossVersionObjectReference `json:"scaleTargetRef"`
+
+	// Behavior defines scaling actions, which are transparently transmitted to HPA and implemented by HPA Controller.
+	// +optional
+	Behavior *autoscalingv2.HorizontalPodAutoscalerBehavior `json:"behavior,omitempty"`
+
+	// Metrics define the metrics used for workload scaling.
+	// +optional
+	Metrics []MetricSpec `json:"metrics,omitempty"`
+
+	// MinReplicas and MaxReplicas are consistent with those in HPA.
+	// +optional
+	MinReplicas *int32 `json:"minReplicas,omitempty"`
+	MaxReplicas int32  `json:"maxReplicas"`
+}
+
+// MetricSpec is used to define a single metric of HPA.
+// Metric is preferred. Only one of Metric and CustomMetric will take effect.
+// The metrics source of Metric is API Server.
+// The metrics source of CustomMetric is the resource portrait.
+type MetricSpec struct {
+	Metric       *autoscalingv2.MetricSpec `json:"metric,omitempty"`
+	CustomMetric *CustomMetricSpec         `json:"customMetric,omitempty"`
+}
+
+// CustomMetricSpec configures metrics derived from resource portraits.
+type CustomMetricSpec struct {
+	// Identify defines the name of the resource metric
+	Identify corev1.ResourceName `json:"identify"`
+
+	// Query defines the metrics query statement.
+	// There are preset templates for resource portraits, so the query statement can be empty.
+	Query string `json:"query,omitempty"`
+
+	// Value represents the threshold, corresponding to the AverageValue in HPA.
+	Value *resource.Quantity `json:"value"`
+}
+
+// TimeBound supports adjusting HPA max/min replicas based on the period and the time within the period.
+type TimeBound struct {
+	// Start and End are the time period.
+	Start  metav1.Time `json:"start,omitempty"`
+	End    metav1.Time `json:"end,omitempty"`
+	Bounds []Bound     `json:"bounds,omitempty"`
+}
+
+// Bound defines the max/min replicas of HPA configured under the specified CronTab.
+type Bound struct {
+	CronTab string `json:"cronTab"`
+
+	MinReplicas *int32 `json:"minReplicas,omitempty"`
+	MaxReplicas *int32 `json:"maxReplicas,omitempty"`
+}
+
+// ScaleStrategyType is the strategy of IHPA to scale workloads.
+type ScaleStrategyType string
+
+const (
+	Preview ScaleStrategyType = "Preview"
+	Auto    ScaleStrategyType = "Auto"
+)
+
+// IntelligentHorizontalPodAutoscalerConditionType are the valid conditions of IntelligentHorizontalPodAutoscaler
+type IntelligentHorizontalPodAutoscalerConditionType string
+
+const (
+	// ServiceProfileDescriptorSynced indicates that the SPD had been synced.
+	ServiceProfileDescriptorSynced IntelligentHorizontalPodAutoscalerConditionType = "SPDSynced"
+	// VirtualWorkloadSynced indicates that the VW had been synced.
+	VirtualWorkloadSynced IntelligentHorizontalPodAutoscalerConditionType = "VWCreated"
+	// HorizontalPodAutoscalerSynced indicates that the HPA had been synced.
+	HorizontalPodAutoscalerSynced IntelligentHorizontalPodAutoscalerConditionType = "HPASynced"
+)
+
+// IntelligentHorizontalPodAutoscalerCondition describes the state of
+// a IntelligentHorizontalPodAutoscaler at a certain point.
+type IntelligentHorizontalPodAutoscalerCondition struct {
+	// type describes the current condition
+	Type IntelligentHorizontalPodAutoscalerConditionType `json:"type"`
+	// status is the status of the condition (True, False, Unknown)
+	Status corev1.ConditionStatus `json:"status"`
+	// lastTransitionTime is the last time the condition transitioned from
+	// one status to another
+	// +optional
+	LastTransitionTime metav1.Time `json:"lastTransitionTime,omitempty"`
+	// reason is the reason for the condition's last transition.
+	// +optional
+	Reason string `json:"reason,omitempty"`
+	// message is a human-readable explanation containing details about
+	// the transition
+	// +optional
+	Message string `json:"message,omitempty"`
+}
+
+// +k8s:deepcopy-gen:interfaces=k8s.io/apimachinery/pkg/runtime.Object
+
+// IntelligentHorizontalPodAutoscalerList is a collection of IntelligentHorizontalPodAutoscaler objects.
+type IntelligentHorizontalPodAutoscalerList struct {
+	metav1.TypeMeta `json:",inline"`
+
+	// Standard list metadata
+	// More info: https://git.k8s.io/community/contributors/devel/api-conventions.md#metadata
+	// +optional
+	metav1.ListMeta `json:"metadata,omitempty"`
+
+	// items is the list of IntelligentHorizontalPodAutoscaler
+	Items []IntelligentHorizontalPodAutoscaler `json:"items"`
+}
+
+// VirtualWorkloadSpec defines the desired state of VirtualWorkload
+type VirtualWorkloadSpec struct {
+	Replicas int32 `json:"replicas"`
+}
+
+// VirtualWorkloadStatus defines the observed state of VirtualWorkload
+type VirtualWorkloadStatus struct {
+	Replicas int32  `json:"replicas"`
+	Selector string `json:"selector"`
+}
+
+// +genclient
+// +k8s:deepcopy-gen:interfaces=k8s.io/apimachinery/pkg/runtime.Object
+//+kubebuilder:object:root=true
+//+kubebuilder:subresource:status
+//+kubebuilder:subresource:scale:specpath=.spec.replicas,statuspath=.status.replicas,selectorpath=.status.selector
+
+// VirtualWorkload is the Schema for the virtualworkloads API
+// VirtualWorkload is used to support IHPA's Preview mode, that is, by providing
+// a virtual workload reference so that scaling will not affect the real workload.
+type VirtualWorkload struct {
+	metav1.TypeMeta   `json:",inline"`
+	metav1.ObjectMeta `json:"metadata,omitempty"`
+
+	Spec   VirtualWorkloadSpec   `json:"spec,omitempty"`
+	Status VirtualWorkloadStatus `json:"status,omitempty"`
+}
+
+// +k8s:deepcopy-gen:interfaces=k8s.io/apimachinery/pkg/runtime.Object
+
+// VirtualWorkloadList contains a list of VirtualWorkload
+type VirtualWorkloadList struct {
+	metav1.TypeMeta `json:",inline"`
+
+	// Standard list metadata
+	// More info: https://git.k8s.io/community/contributors/devel/api-conventions.md#metadata
+	// +optional
+	metav1.ListMeta `json:"metadata,omitempty"`
+
+	// items is the list of VirtualWorkload
+	Items []VirtualWorkload `json:"items"`
+}

pkg/apis/autoscaling/v1alpha2/register.go

@@ -32,7 +32,9 @@ var SchemeGroupVersion = schema.GroupVersion{Group: GroupName, Version: "v1alpha
 
 // ResourceName const is used to construct standard gvr
 const (
-	ResourceNameKatalystVPA = "katalystverticalpodautoscalers"
+	ResourceNameKatalystVPA     = "katalystverticalpodautoscalers"
+	ResourceNameIHPA            = "intelligenthorizontalpodautoscalers"
+	ResourceNameVirtualWorkload = "virtualworkloads"
 )
 
 // Resource takes an unqualified resource and returns a Group qualified GroupResource
@@ -52,6 +54,10 @@ func addKnownTypes(scheme *runtime.Scheme) error {
 	scheme.AddKnownTypes(SchemeGroupVersion,
 		&KatalystVerticalPodAutoscaler{},
 		&KatalystVerticalPodAutoscalerList{},
+		&IntelligentHorizontalPodAutoscaler{},
+		&IntelligentHorizontalPodAutoscalerList{},
+		&VirtualWorkload{},
+		&VirtualWorkloadList{},
 	)
 
 	metav1.AddToGroupVersion(scheme, SchemeGroupVersion)