feat: add HPA autoscaling for InferenceService

api/v1alpha1/inferenceservice_types.go

@@ -37,6 +37,14 @@ type InferenceServiceSpec struct {
 	// +optional
 	Replicas *int32 `json:"replicas,omitempty"`
 
+	// Autoscaling configures horizontal pod autoscaling for the inference service.
+	// When set, the controller creates and manages an HPA resource targeting the
+	// inference Deployment. Requires Prometheus Adapter for custom metrics.
+	// Mutually exclusive with manual replica management: when autoscaling is enabled,
+	// the Replicas field serves as the initial replica count only.
+	// +optional
+	Autoscaling *AutoscalingSpec `json:"autoscaling,omitempty"`
+
 	// Image is the container image for the llama.cpp runtime
 	// +kubebuilder:default="ghcr.io/ggml-org/llama.cpp:server"
 	// +optional
@@ -172,6 +180,44 @@ type InferenceResourceRequirements struct {
 	GPUMemory string `json:"gpuMemory,omitempty"`
 }
 
+// AutoscalingSpec configures Horizontal Pod Autoscaler for the inference service.
+type AutoscalingSpec struct {
+	// MinReplicas is the lower limit for the number of replicas.
+	// +kubebuilder:validation:Minimum=1
+	// +kubebuilder:validation:Maximum=10
+	// +kubebuilder:default=1
+	// +optional
+	MinReplicas *int32 `json:"minReplicas,omitempty"`
+
+	// MaxReplicas is the upper limit for the number of replicas.
+	// +kubebuilder:validation:Minimum=1
+	// +kubebuilder:validation:Maximum=100
+	MaxReplicas int32 `json:"maxReplicas"`
+
+	// Metrics defines the scaling metrics and target values.
+	// If empty, defaults to llamacpp:requests_processing with target average value of 2.
+	// +optional
+	Metrics []MetricSpec `json:"metrics,omitempty"`
+}
+
+// MetricSpec defines a single metric for HPA scaling.
+type MetricSpec struct {
+	// Type is the metric source type.
+	// +kubebuilder:validation:Enum=Pods;Resource
+	Type string `json:"type"`
+
+	// Name is the metric name (e.g., llamacpp:requests_processing).
+	Name string `json:"name"`
+
+	// TargetAverageValue is the target per-pod average for Pods-type metrics.
+	// +optional
+	TargetAverageValue *string `json:"targetAverageValue,omitempty"`
+
+	// TargetAverageUtilization is the target utilization percentage for Resource-type metrics.
+	// +optional
+	TargetAverageUtilization *int32 `json:"targetAverageUtilization,omitempty"`
+}
+
 // InferenceServiceStatus defines the observed state of InferenceService.
 type InferenceServiceStatus struct {
 	// Phase represents the current lifecycle phase (Pending, Creating, Ready, Failed)

charts/llmkube/templates/clusterrole.yaml

@@ -49,6 +49,18 @@ rules:
   - patch
   - update
   - watch
+- apiGroups:
+  - autoscaling
+  resources:
+  - horizontalpodautoscalers
+  verbs:
+  - create
+  - delete
+  - get
+  - list
+  - patch
+  - update
+  - watch
 - apiGroups:
   - inference.llmkube.dev
   resources:

config/crd/bases/inference.llmkube.dev_inferenceservices.yaml

@@ -69,6 +69,62 @@ spec:
           spec:
             description: spec defines the desired state of InferenceService
             properties:
+              autoscaling:
+                description: |-
+                  Autoscaling configures horizontal pod autoscaling for the inference service.
+                  When set, the controller creates and manages an HPA resource targeting the
+                  inference Deployment. Requires Prometheus Adapter for custom metrics.
+                  Mutually exclusive with manual replica management: when autoscaling is enabled,
+                  the Replicas field serves as the initial replica count only.
+                properties:
+                  maxReplicas:
+                    description: MaxReplicas is the upper limit for the number of
+                      replicas.
+                    format: int32
+                    maximum: 100
+                    minimum: 1
+                    type: integer
+                  metrics:
+                    description: |-
+                      Metrics defines the scaling metrics and target values.
+                      If empty, defaults to llamacpp:requests_processing with target average value of 2.
+                    items:
+                      description: MetricSpec defines a single metric for HPA scaling.
+                      properties:
+                        name:
+                          description: Name is the metric name (e.g., llamacpp:requests_processing).
+                          type: string
+                        targetAverageUtilization:
+                          description: TargetAverageUtilization is the target utilization
+                            percentage for Resource-type metrics.
+                          format: int32
+                          type: integer
+                        targetAverageValue:
+                          description: TargetAverageValue is the target per-pod average
+                            for Pods-type metrics.
+                          type: string
+                        type:
+                          description: Type is the metric source type.
+                          enum:
+                          - Pods
+                          - Resource
+                          type: string
+                      required:
+                      - name
+                      - type
+                      type: object
+                    type: array
+                  minReplicas:
+                    default: 1
+                    description: MinReplicas is the lower limit for the number of
+                      replicas.
+                    format: int32
+                    maximum: 10
+                    minimum: 1
+                    type: integer
+                required:
+                - maxReplicas
+                type: object
               cacheTypeK:
                 description: |-
                   CacheTypeK sets the KV cache quantization type for keys.

config/rbac/role.yaml

@@ -47,6 +47,18 @@ rules:
   - patch
   - update
   - watch
+- apiGroups:
+  - autoscaling
+  resources:
+  - horizontalpodautoscalers
+  verbs:
+  - create
+  - delete
+  - get
+  - list
+  - patch
+  - update
+  - watch
 - apiGroups:
   - inference.llmkube.dev
   resources: