scalablestats

README

ScalableStats

ScalableStats is a library to allow collection and processing of metrics with O(1) constant time and space guarantees; this makes it suitable for frequently-sampled metrics such as temperature.

Guarantees are provided by forming a histogram based on apriori knowledge of the data, for example sampled temperatures will lie in the range of 10 to 80'C, quantized to 2'C, ie 35 bins.

Arbitrary percentiles can be extracted from the resultant histogram.

Example

h := NewLinearHistogram(10, 80, 35)

for i := 0; i < 100000; i++ {
    h.Store(rand.Float32() * 80)
}

fmt.Printf("P95=%v\n", h.Percentile(95))
h.Fprint(os.Stdout)
h.CLear()

Documentation

Index

• type Histogram
  • func NewLinearHistogram(min, max, bins int) *Histogram
  • func (h *Histogram) Clear()
  • func (h *Histogram) Fprint(f *os.File)
  • func (h *Histogram) Percentile(threshold int) float32
  • func (h *Histogram) Store(val float32)

Types

type Histogram

type Histogram struct {
	Thresholds []float32
	Counts     []int
}

func NewLinearHistogram

func NewLinearHistogram(min, max, bins int) *Histogram

func (*Histogram) Clear

func (h *Histogram) Clear()

func (*Histogram) Fprint

func (h *Histogram) Fprint(f *os.File)

func (*Histogram) Percentile

func (h *Histogram) Percentile(threshold int) float32

func (*Histogram) Store

func (h *Histogram) Store(val float32)