scaler

type ApaAutoscaler struct {
	Status ScaleResult
	// contains filtered or unexported fields
}

type ApaScalingContext struct {
	scalingcontext.BaseScalingContext

	// The two following attributes are specific to APA.
	// UpFluctuationTolerance represents the threshold before scaling up,
	// which means no scaling up will occur unless the currentMetricValue exceeds the TargetValue by more than UpFluctuationTolerance
	UpFluctuationTolerance float64
	// UpFluctuationTolerance represents the threshold before scaling down,
	// which means no scaling down will occur unless the currentMetricValue is less than the TargetValue by more than UpFluctuationTolerance
	DownFluctuationTolerance float64
	// metric window length
	Window time.Duration
}

type KpaAutoscaler struct {
	Status *ScaleResult
	// contains filtered or unexported fields
}

type KpaScalingContext struct {
	scalingcontext.BaseScalingContext

	// The burst capacity that user wants to maintain without queuing at the POD level.
	// Note, that queueing still might happen due to the non-ideal load balancing.
	TargetBurstCapacity float64
	// ActivationScale is the minimum, non-zero value that a service should scale to.
	// For example, if ActivationScale = 2, when a service scaled from zero it would
	// scale up two replicas in this case. In essence, this allows one to set both a
	// min-scale value while also preserving the ability to scale to zero.
	// ActivationScale must be >= 1.
	ActivationScale int32

	// TODO: Note that the following attributes are specific to Knative; but we retain them here temporarily.
	// PanicThreshold is the threshold at which panic mode is entered. It represents
	// a factor of the currently observed load over the panic window over the ready
	// pods. I.e. if this is 2, panic mode will be entered if the observed metric
	// is twice as high as the current population can handle.
	PanicThreshold float64
	// StableWindow is needed to determine when to exit panic mode.
	StableWindow time.Duration
	// PanicWindow is needed to determine when to exit panic mode.
	PanicWindow time.Duration
	// ScaleDownDelay is the time that must pass at reduced concurrency before a
	// scale-down decision is applied.
	ScaleDownDelay time.Duration
}

type ScaleResult struct {
	// DesiredPodCount is the number of pods Autoscaler suggests for the revision.
	DesiredPodCount int32
	// ExcessBurstCapacity is computed headroom of the revision taking into
	// the account target burst capacity.
	ExcessBurstCapacity int32
	// ScaleValid specifies whether this scale result is valid, i.e. whether
	// Autoscaler had all the necessary information to compute a suggestion.
	ScaleValid bool
}

type Scaler interface {
	// UpdateScaleTargetMetrics updates the current state of metrics used to determine scaling actions.
	// It processes the latest metrics for a given scaling target (identified by metricKey) and stores
	// these values for later use during scaling decisions.
	//
	// Parameters:
	// - ctx: The context used for managing request-scoped values, cancellation, and deadlines.
	// - metricKey: A unique identifier for the scaling target's metrics (e.g., CPU, memory, or QPS) that
	//   is used to correlate metrics with the appropriate scaling logic.
	// - now: The current time at which the metrics are being processed. This timestamp helps track
	//   when the last metric update occurred and can be used to calculate time-based scaling actions.
	//
	// This method ensures that the autoscaler has up-to-date metrics before making any scaling decisions.
	UpdateScaleTargetMetrics(ctx context.Context, metricKey metrics.NamespaceNameMetric, source autoscalingv1alpha1.MetricSource, pods []corev1.Pod, now time.Time) error

	// UpdateSourceMetrics updates the current state of metrics used to determine scaling actions.
	// It processes the latest metrics for a metrics source and stores
	// these values for later use during scaling decisions.
	//
	// Parameters:
	// - ctx: The context used for managing request-scoped values, cancellation, and deadlines.
	// - metricKey: A unique identifier for the scaling target's metrics (e.g., CPU, memory, or QPS) that
	//   is used to correlate metrics with the appropriate scaling logic.
	// - source: The MetricSource object containing the desired scaling configuration and current state.
	// - now: The current time at which the metrics are being processed. This timestamp helps track
	//   when the last metric update occurred and can be used to calculate time-based scaling actions.
	//
	// This method ensures that the autoscaler has up-to-date metrics before making any scaling decisions.
	UpdateSourceMetrics(ctx context.Context, metricKey metrics.NamespaceNameMetric, source autoscalingv1alpha1.MetricSource, now time.Time) error

	// Scale calculates the necessary scaling action based on observed metrics
	// and the current time. This is the core logic of the autoscaler.
	//
	// Parameters:
	// originalReadyPodsCount - the current number of ready pods.
	// metricKey - a unique key to identify the metric for scaling.
	// now - the current time, used to decide if scaling actions are needed based on timing rules or delays.
	//
	// Returns:
	// ScaleResult - contains the recommended number of pods to scale up or down.
	//
	// For reference: see the implementation in KpaAutoscaler.Scale.
	Scale(originalReadyPodsCount int, metricKey metrics.NamespaceNameMetric, now time.Time) ScaleResult

	// UpdateScalingContext updates the internal scaling context for a given PodAutoscaler (PA) instance.
	// It extracts necessary information from the provided PodAutoscaler resource, such as current
	// metrics, scaling parameters, and other relevant data to refresh the scaling context.
	//
	// Parameters:
	// - pa: The PodAutoscaler resource containing the desired scaling configuration and current state.
	//
	// Returns:
	// - error: If the context update fails due to invalid input or configuration issues, it returns an error.
	//
	// This method ensures that the internal scaling context is always in sync with the latest state
	// and configuration of the target PodAutoscaler, allowing accurate scaling decisions.
	UpdateScalingContext(pa autoscalingv1alpha1.PodAutoscaler) error

	// GetScalingContext retrieves the current scaling context used for making scaling decisions.
	// This method returns a pointer to the ScalingContext, which contains essential data like
	// target values, current metrics, and scaling tolerances.
	//
	// Returns:
	// - *common.ScalingContext: A pointer to the ScalingContext instance containing the relevant
	//   data for autoscaling logic.
	//
	// This method provides access to the scaling context for external components or logic that
	// need to read or adjust the current scaling parameters.
	GetScalingContext() common.ScalingContext
}