Documentation
¶
Index ¶
- func AddMetricsToOptStatus(ctx context.Context, opt *llmdVariantAutoscalingV1alpha1.VariantAutoscaling, ...) (llmdVariantAutoscalingV1alpha1.Allocation, error)
- func CollectInventoryK8S(ctx context.Context, r interface{}) (map[string]map[string]AcceleratorModelInfo, error)
- func FixValue(x *float64)
- type AcceleratorModelInfo
- type MetricKV
- type MetricsValidationResult
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
func AddMetricsToOptStatus ¶
func AddMetricsToOptStatus(ctx context.Context, opt *llmdVariantAutoscalingV1alpha1.VariantAutoscaling, deployment appsv1.Deployment, acceleratorCostVal float64, promAPI promv1.API) (llmdVariantAutoscalingV1alpha1.Allocation, error)
func CollectInventoryK8S ¶
func CollectInventoryK8S(ctx context.Context, r interface{}) (map[string]map[string]AcceleratorModelInfo, error)
CollectInventoryK8S is a stub for future limited mode support. Currently returns empty inventory as WVA operates in unlimited mode.
Types ¶
type AcceleratorModelInfo ¶
type MetricsValidationResult ¶ added in v0.0.2
MetricsValidationResult contains the result of metrics availability check
func ValidateMetricsAvailability ¶ added in v0.0.2
func ValidateMetricsAvailability(ctx context.Context, promAPI promv1.API, modelName, namespace string) MetricsValidationResult
ValidateMetricsAvailability checks if vLLM metrics are available for the given model and namespace Returns a validation result with details about metric availability
Click to show internal directories.
Click to hide internal directories.