README
¶
Autoscaling Algorithms
This package provides various scaling algorithms for Pod Autoscaling, including implementations for
- APA (Adaptive Pod Autoscaler),
- KPA (KNative Pod Autoscaler),
- HPA (Horizontal Pod Autoscaler), and more.
These algorithms are designed to dynamically compute the desired number of replicas based on current pod usage and scaling specifications, optimizing resource usage and ensuring high availability and performance for workloads.
ScalingAlgorithm Interface is a common interface for all scaling algorithms, requiring the implementation of the ComputeTargetReplicas method,
which calculates the number of replicas based on current metrics and scaling specifications.
type ScalingAlgorithm interface {
ComputeTargetReplicas(currentPodCount float64, context ScalingContext) int32
}
Documentation
¶
Index ¶
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
This section is empty.
Types ¶
type ApaScalingAlgorithm ¶
type ApaScalingAlgorithm struct{}
func (*ApaScalingAlgorithm) ComputeTargetReplicas ¶
func (a *ApaScalingAlgorithm) ComputeTargetReplicas(currentPodCount float64, context common.ScalingContext) int32
ComputeTargetReplicas - Apa's algorithm references and enhances the algorithm in the following paper: Huo, Qizheng, et al. "High Concurrency Response Strategy based on Kubernetes Horizontal Pod Autoscaler." Journal of Physics: Conference Series. Vol. 2451. No. 1. IOP Publishing, 2023.
type HpaScalingAlgorithm ¶
type HpaScalingAlgorithm struct{}
HpaScalingAlgorithm can be used by any scaler without customized algorithms
func (*HpaScalingAlgorithm) ComputeTargetReplicas ¶
func (a *HpaScalingAlgorithm) ComputeTargetReplicas(currentPodCount float64, context common.ScalingContext) int32
type KpaScalingAlgorithm ¶
type KpaScalingAlgorithm struct{}
func (*KpaScalingAlgorithm) ComputeTargetReplicas ¶
func (a *KpaScalingAlgorithm) ComputeTargetReplicas(currentPodCount float64, context common.ScalingContext) int32
type ScalingAlgorithm ¶
type ScalingAlgorithm interface {
// ComputeTargetReplicas calculates the number of replicas needed based on current metrics
// and the provided scaling specifications.
//
// Parameters:
// currentPodCount - the current number of ready pods
// context - an interface that provides access to scaling parameters like target values and tolerances
//
// Returns:
// int32 - the calculated target number of replicas
ComputeTargetReplicas(currentPodCount float64, context common.ScalingContext) int32
}
Source Files