Package com.twitter.common.stats

Interface Summary
Histogram An interface for Histogram
Stat<T> A stat that may only be read, no method calls will cause any internal state changes.
StatRegistry A registry of stats.
StatsProvider A minimal interface to a Stats repository.
StatsProvider.RequestTimer A stat for tracking service requests.
TimeSeries A time series of values.
TimeSeriesRepository A repository for time series data.
 

Class Summary
ApproximateHistogram Implements Histogram structure for computing approximate quantiles.
CounterMap<K> A map from a key type to integers.
CounterMapWithTopKey<K> Same as CounterMap, but also keeps track of the item with the highest count.
Elapsed A stat that exports the amount of time since it was last reset.
Entropy<T> Calculate the entropy of a discrete distribution of .
JvmStats Convenience class to export statistics about the JVM.
MovingAverage<T extends Number> Function to compute the moving average of a time series.
MovingWindowDelta<T extends Number> Delta over the most recent k sample periods.
NumericStatExporter Stat exporter that extracts numeric Stats from the Stats system, and exports them via a caller-defined sink.
Percentile<T extends Number & Comparable<T>> A stats tracker to export percentiles of inputs based on a sampling rate.
PipelineStats Tracks the latency of different pipeline stages in a process.
Precision Precision expresses the maximum epsilon tolerated for a typical size of input e.g.: Precision(0.01, 1000) express that we tolerate a error of 1% for 1000 entries it means that max difference between the real quantile and the estimate one is error = 0.01*1000 = 10 For an entry like (1 to 1000), q(0.5) will be [490 <= x <= 510] (real q(0.5) = 500)
PrintableHistogram  
Rate<T extends Number> Function to compute a windowed per-second rate of a value.
Rate.Builder<T extends Number>  
Ratio Function to compute the ratio of two time series.
RequestStats A class to represent the statistics associated with a client connection to an external service.
ReservoirSampler<T> An in memory implementation of Reservoir Sampling for sampling from a population.
SampledStat<T extends Number> A convenience class to perform the basic tasks needed for a RecordingStat except the actual value calculation.
Significance Calculate significance scores between an observed amount and an expected amount.
SlidingStats Tracks event statistics over a sliding window of time.
StatImpl<T> A convenience class to not require stat implementers to implement StatImpl.getName().
Statistics A simple class to keep running statistics that require O(1) storage.
Stats Manages Stats that should be exported for monitoring.
TimeSeriesRepositoryImpl A simple in-memory repository for exported variables.