Documentation
¶
Overview ¶
Package algorithm provides scaling algorithms for different autoscaling strategies. Each strategy (KPA, APA, HPA) has its own algorithm implementation that computes target replicas based on metrics and scaling context.
Index ¶
Constants ¶
const ( MinScaleRate = 1.0 DefaultTolerance = 0.0 )
Variables ¶
This section is empty.
Functions ¶
This section is empty.
Types ¶
type APAAlgorithm ¶
type APAAlgorithm struct{}
APAAlgorithm implements Application-specific Pod Autoscaling This is a stateless struct that can be safely reused across goroutines
func (*APAAlgorithm) ComputeRecommendation ¶
func (a *APAAlgorithm) ComputeRecommendation(ctx context.Context, request ScalingRequest) (*ScalingRecommendation, error)
ComputeRecommendation calculates desired replica count for APA strategy
func (*APAAlgorithm) GetAlgorithmType ¶
func (a *APAAlgorithm) GetAlgorithmType() string
GetAlgorithmType returns the algorithm type
type HPAAlgorithm ¶
type HPAAlgorithm struct{}
HPAAlgorithm is a placeholder for HPA strategy (actual HPA is handled by K8s HPA resources) This is a stateless struct that can be safely reused across goroutines
func (*HPAAlgorithm) ComputeRecommendation ¶
func (a *HPAAlgorithm) ComputeRecommendation(ctx context.Context, request ScalingRequest) (*ScalingRecommendation, error)
ComputeRecommendation for HPA just returns current replicas as HPA is managed by K8s
func (*HPAAlgorithm) GetAlgorithmType ¶
func (a *HPAAlgorithm) GetAlgorithmType() string
GetAlgorithmType returns the algorithm type
type KPAAlgorithm ¶
type KPAAlgorithm struct{}
KPAAlgorithm implements Knative-style Pod Autoscaling with panic and stable windows This is a stateless struct that can be safely reused across goroutines
func (*KPAAlgorithm) ComputeRecommendation ¶
func (a *KPAAlgorithm) ComputeRecommendation(ctx context.Context, request ScalingRequest) (*ScalingRecommendation, error)
ComputeRecommendation calculates desired replica count for KPA strategy
func (*KPAAlgorithm) GetAlgorithmType ¶
func (a *KPAAlgorithm) GetAlgorithmType() string
GetAlgorithmType returns the algorithm type
type ScalingAlgorithm ¶
type ScalingAlgorithm interface {
// ComputeRecommendation calculates desired replica count based on aggregated metrics
// All configuration is passed via the request to keep the algorithm stateless
ComputeRecommendation(ctx context.Context, request ScalingRequest) (*ScalingRecommendation, error)
// GetAlgorithmType returns the algorithm type (kpa, apa, hpa)
GetAlgorithmType() string
}
ScalingAlgorithm computes target replicas based on metrics and context Implementations must be stateless and thread-safe for concurrent use
func NewScalingAlgorithm ¶
func NewScalingAlgorithm(strategy autoscalingv1alpha1.ScalingStrategyType) ScalingAlgorithm
NewScalingAlgorithm creates a stateless algorithm instance for the given strategy Instances can be safely cached and reused across goroutines
type ScalingRecommendation ¶
type ScalingRecommendation struct {
DesiredReplicas int32
Confidence float64
Reason string
Algorithm string
ScaleValid bool
Metadata map[string]interface{}
}
ScalingRecommendation contains the scaling decision
type ScalingRequest ¶
type ScalingRequest struct {
Target types.ScaleTarget
CurrentReplicas int32
AggregatedMetrics *types.AggregatedMetrics
ScalingContext scalingctx.ScalingContext
Timestamp time.Time
}
ScalingRequest contains all data needed for scaling decision
Source Files