MakeBucketsForError compute all the bucket values from 1 until we run out of positive 64-bit ints for the given an error (+/-).

NewBiasedHistogram returns a histogram that uses an exponentially decaying sample of 1028 elements, which offers a 99.9% confidence level with a 5% margin of error assuming a normal distribution, and an alpha factor of 0.015, which heavily biases the sample to the past 5 minutes of measurements.

NewBucketedHistogram returns a histogram that uses a fixed set of buckets for ranges of values.

NewCounter returns a counter implemented as an atomic uint64.

NewDefaultBucketedHistogram returns a bucketed histogram with an error of 5%.

NewDistribution returns a new instance of a Distribution.

NewEWMA returns a new exponentially-weighte moving average.

NewExponentiallyDecayingSample returns an exponentially-decaying random sample of values.

NewExponentiallyDecayingSampleWithCustomTime returns an exponentially-decaying random sample of values using a custom time function.

NewMeter returns a new instance of Meter.

NewMunroPatersonHistogram returns an implemenation of the Munro-Paterson approximate histogram algorithm adapted from: https://github.com/twitter/commons/blob/master/src/java/com/twitter/common/stats/ApproximateHistogram.java http://szl.googlecode.com/svn-history/r36/trunk/src/emitters/szlquantile.cc.

NewUnbiasedHistogram returns a histogram that uses a uniform sample of 1028 elements, which offers a 99.9% confidence level with a 5% margin of error assuming a normal distribution.

NewUniformSample returns a sample randomly selects from a stream.

M15Alpha represents 15 minutes at a 5 second interval.

M1Alpha represents 1 minute at a 5 second interval.

M5Alpha represents 5 minutes at a 5 second interval.

Distribution tracks the min, max, sum, count, and variance/stddev of a set of values.

EWMA is an exponentially-weighted moving average.

Meter is the combination of three EWMA metrics: 1 min, 5 min, and 15 min.

Precision expresses the maximum epsilon tolerated for a typical size of input.

Counter is the interface for a counter metric.

FloatGaugeFunc is used for stats reporting to identify the value as a floating point gauge.