Package com.tnsai.quality.sampling

Class AdaptiveSamplingPolicy

java.lang.Object

com.tnsai.quality.sampling.AdaptiveSamplingPolicy

All Implemented Interfaces:

EventSamplingPolicy

public final class AdaptiveSamplingPolicy
extends Object
implements EventSamplingPolicy

Backpressure-aware EventSamplingPolicy (issue #82). Wraps a delegate policy and reduces effective sample rate when an external AdaptiveSamplingPolicy.BackpressureSignal reports that the publisher chain is overloaded.

The signal is intentionally minimal — a double in [0.0, 1.0] representing "fraction of capacity in use". The publisher / aggregator owns measurement (queue depth, ack lag, Resilience4j RateLimiter saturation). This class consumes that signal:

• utilisation < lowWaterMark — pass through to delegate
• utilisation >= highWaterMark — drop everything below criticalLevel
• otherwise — linear-interpolate the delegate's SamplingDecision.Capture into a downsampled capture

Gives operators a way to gracefully degrade event volume under sustained load without losing critical events.

Since:

0.10.0 (issue #82 closeout)

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static interface	AdaptiveSamplingPolicy.BackpressureSignal	Externally-supplied utilisation signal.

Constructor Summary

Constructors

Constructor	Description
AdaptiveSamplingPolicy(EventSamplingPolicy delegate, AdaptiveSamplingPolicy.BackpressureSignal signal)	Defaults: low=0.5, high=0.9, criticalLevel=ERROR.
AdaptiveSamplingPolicy(EventSamplingPolicy delegate, AdaptiveSamplingPolicy.BackpressureSignal signal, double lowWaterMark, double highWaterMark, com.tnsai.observability.events.EventLevel criticalLevel)

Method Summary

Modifier and Type	Method	Description
long	droppedDueToPressureCount()	How many events did pressure-induced scaling drop?
SamplingDecision	sample(SamplingInput input)	Inspect an event's metadata and return whether to capture, drop, or capture-with-downsampling marker.

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details

AdaptiveSamplingPolicy

public AdaptiveSamplingPolicy(EventSamplingPolicy delegate,
 AdaptiveSamplingPolicy.BackpressureSignal signal,
 double lowWaterMark,
 double highWaterMark,
 com.tnsai.observability.events.EventLevel criticalLevel)

AdaptiveSamplingPolicy

public AdaptiveSamplingPolicy(EventSamplingPolicy delegate,
 AdaptiveSamplingPolicy.BackpressureSignal signal)

Defaults: low=0.5, high=0.9, criticalLevel=ERROR. Below 50% load the delegate runs unchanged; between 50-90% load the capture rate scales down linearly; above 90% only ERROR-and-above survive.

Method Details

sample

public SamplingDecision sample(SamplingInput input)

Description copied from interface: EventSamplingPolicy

Inspect an event's metadata and return whether to capture, drop, or capture-with-downsampling marker.

Specified by:

sample in interface EventSamplingPolicy

droppedDueToPressureCount

public long droppedDueToPressureCount()

How many events did pressure-induced scaling drop? Monitoring counter.