aggregation

Documentation

Index

• Variables
• type Aggregation
• type Buckets
• type Count
• type Histogram
• type Kind
  • func (k Kind) String() string
• type LastValue
• type Max
• type Min
• type MinMaxSumCount
• type Point
• type Points
• type Sum

Variables

var (
	ErrNegativeInput    = fmt.Errorf("negative value is out of range for this instrument")
	ErrNaNInput         = fmt.Errorf("NaN value is an invalid input")
	ErrInconsistentType = fmt.Errorf("inconsistent aggregator types")
	ErrNoSubtraction    = fmt.Errorf("aggregator does not subtract")

	// ErrNoData is returned when (due to a race with collection)
	// the Aggregator is check-pointed before the first value is set.
	// The aggregator should simply be skipped in this case.
	ErrNoData = fmt.Errorf("no data collected by this aggregator")
)

Types

type Aggregation

type Aggregation interface {
	// Kind returns a short identifying string to identify
	// the Aggregator that was used to produce the
	// Aggregation (e.g., "Sum").
	Kind() Kind
}

Aggregation is an interface returned by the Aggregator containing an interval of metric data.

type Buckets

type Buckets struct {
	// Boundaries are floating point numbers, even when
	// aggregating integers.
	Boundaries []float64

	// Counts holds the count in each bucket.
	Counts []uint64
}

Buckets represents histogram buckets boundaries and counts.

For a Histogram with N defined boundaries, e.g, [x, y, z]. There are N+1 counts: [-inf, x), [x, y), [y, z), [z, +inf]

type Count

type Count interface {
	Aggregation
	Count() (uint64, error)
}

Count returns the number of values that were aggregated.

type Histogram

type Histogram interface {
	Aggregation
	Count() (uint64, error)
	Sum() (number.Number, error)
	Histogram() (Buckets, error)
}

Histogram returns the count of events in pre-determined buckets.

type Kind

type Kind string

Kind is a short name for the Aggregator that produces an Aggregation, used for descriptive purpose only. Kind is a string to allow user-defined Aggregators.

When deciding how to handle an Aggregation, Exporters are encouraged to decide based on conversion to the above interfaces based on strength, not on Kind value, when deciding how to expose metric data. This enables user-supplied Aggregators to replace builtin Aggregators.

For example, test for a Distribution before testing for a MinMaxSumCount, test for a Histogram before testing for a Sum, and so on.

const (
	SumKind            Kind = "Sum"
	MinMaxSumCountKind Kind = "MinMaxSumCount"
	HistogramKind      Kind = "Histogram"
	LastValueKind      Kind = "Lastvalue"
	ExactKind          Kind = "Exact"
)

func (Kind) String

func (k Kind) String() string

String returns the string value of Kind.

type LastValue

type LastValue interface {
	Aggregation
	LastValue() (number.Number, time.Time, error)
}

LastValue returns the latest value that was aggregated.

type Max

type Max interface {
	Aggregation
	Max() (number.Number, error)
}

Max returns the maximum value over the set of values that were aggregated.

type Min

type Min interface {
	Aggregation
	Min() (number.Number, error)
}

Min returns the minimum value over the set of values that were aggregated.

type MinMaxSumCount

type MinMaxSumCount interface {
	Aggregation
	Min() (number.Number, error)
	Max() (number.Number, error)
	Sum() (number.Number, error)
	Count() (uint64, error)
}

MinMaxSumCount supports the Min, Max, Sum, and Count interfaces.

type Point

type Point struct {
	number.Number
	time.Time
}

Point is a raw data point, consisting of a number and value.

type Points

type Points interface {
	Aggregation

	// Points returns points in the order they were
	// recorded.  Points are approximately ordered by
	// timestamp, but this is not guaranteed.
	Points() ([]Point, error)
}

Points returns the raw values that were aggregated.

type Sum

type Sum interface {
	Aggregation
	Sum() (number.Number, error)
}

Sum returns an aggregated sum.