actionablecluster

Documentation

Index

• Constants
• func OnEmptyCluster(context *context.AutoscalingContext, status string, emitEvent bool)
• type ActionableClusterProcessor
  • func NewCustomActionableClusterProcessor(startTime time.Time, nodeNotReadyTimeout time.Duration) ActionableClusterProcessor
  • func NewDefaultActionableClusterProcessor() ActionableClusterProcessor
• type EmptyClusterProcessor
  • func (e *EmptyClusterProcessor) CleanUp()
  • func (e *EmptyClusterProcessor) ShouldAbort(context *context.AutoscalingContext, allNodes []*apiv1.Node, ...) (bool, errors.AutoscalerError)

Constants

const NodesNotReadyAfterStartTimeout = 10 * time.Minute

NodesNotReadyAfterStartTimeout How long should Cluster Autoscaler wait for nodes to become ready after start.

Functions

func OnEmptyCluster

func OnEmptyCluster(context *context.AutoscalingContext, status string, emitEvent bool)

OnEmptyCluster runs actions if the cluster is empty

Types

type ActionableClusterProcessor

type ActionableClusterProcessor interface {
	// ShouldAbort is the func that will return where the cluster is in an actionable state or not
	ShouldAbort(context *context.AutoscalingContext, allNodes []*apiv1.Node, readyNodes []*apiv1.Node, currentTime time.Time) (abortLoop bool, err errors.AutoscalerError)
	CleanUp()
}

ActionableClusterProcessor defines interface for whether action can be taken on the cluster or not

func NewCustomActionableClusterProcessor

func NewCustomActionableClusterProcessor(startTime time.Time, nodeNotReadyTimeout time.Duration) ActionableClusterProcessor

NewCustomActionableClusterProcessor returns a new instance with custom values

func NewDefaultActionableClusterProcessor

func NewDefaultActionableClusterProcessor() ActionableClusterProcessor

NewDefaultActionableClusterProcessor returns a new Processor instance

type EmptyClusterProcessor

type EmptyClusterProcessor struct {
	// contains filtered or unexported fields
}

EmptyClusterProcessor implements ActionableClusterProcessor by using ScaleUpFromZero flag and other node readiness conditions to check whether the cluster is actionable or not

func (*EmptyClusterProcessor) CleanUp

func (e *EmptyClusterProcessor) CleanUp()

CleanUp cleans up the Processor

func (*EmptyClusterProcessor) ShouldAbort

func (e *EmptyClusterProcessor) ShouldAbort(context *context.AutoscalingContext, allNodes []*apiv1.Node, readyNodes []*apiv1.Node, currentTime time.Time) (bool, errors.AutoscalerError)

ShouldAbort give the decision on whether CA can act on the cluster