types

Documentation

Overview

Package types contains miscellaneous structs and interfaces used throughout tsmon. They live here to avoid circular dependencies between other packages.

Index

• Constants
• type Cell
• type CellData
• type DistributionMetric
• type Metric
• type MetricDataUnits
  • func (units MetricDataUnits) IsSpecified() bool
• type MetricInfo
• type MetricMetadata
• type Target
• type ValueType
  • func (v ValueType) IsCumulative() bool
  • func (v ValueType) String() string

Constants

const (
	Unknown      = ""
	Seconds      = "s"
	Milliseconds = "ms"
	Microseconds = "us"
	Nanoseconds  = "ns"
	Bits         = "B"
	Bytes        = "By"

	Kilobytes = "kBy"  // 1000 bytes (not 1024).
	Megabytes = "MBy"  // 1e6 (1,000,000) bytes.
	Gigabytes = "GBy"  // 1e9 (1,000,000,000) bytes.
	Kibibytes = "kiBy" // 1024 bytes.
	Mebibytes = "MiBy" // 1024^2 (1,048,576) bytes.
	Gibibytes = "GiBy" // 1024^3 (1,073,741,824) bytes.

	// * Extended Units
	AmpUnit           = "A"
	MilliampUnit      = "mA"
	DegreeCelsiusUnit = "Cel"
)

Units of metrics data.

Types

type Cell

type Cell struct {
	MetricInfo
	MetricMetadata
	CellData
}

Cell is the smallest unit of data recorded by tsmon. Metrics can be thought of as multi-dimensional maps (with fields defining the dimensions) - a Cell is one value in that map, with information about its fields and its type.

type CellData

type CellData struct {
	FieldVals []interface{}
	Target    Target
	ResetTime time.Time
	Value     interface{}
}

CellData contains the value of a single cell.

type DistributionMetric

type DistributionMetric interface {
	Metric

	Bucketer() *distribution.Bucketer
}

DistributionMetric is the low-level interface provided by all distribution metrics. It has a Bucketer which is responsible for assigning buckets to samples. Concrete types are defined in the "metrics" package.

type Metric

type Metric interface {
	Info() MetricInfo
	Metadata() MetricMetadata

	// SetFixedResetTime overrides the reset time for this metric.  Usually cells
	// take the current time when they're first assigned a value, but it can be
	// useful to override the reset time when tracking an external counter.
	SetFixedResetTime(t time.Time)
}

Metric is the low-level interface provided by all metrics. Concrete types are defined in the "metrics" package.

type MetricDataUnits

type MetricDataUnits string

MetricDataUnits are enums for the units of metrics data.

func (MetricDataUnits) IsSpecified

func (units MetricDataUnits) IsSpecified() bool

IsSpecified returns true if a unit annotation has been specified for a given metric.

type MetricInfo

type MetricInfo struct {
	Name        string
	Description string
	Fields      []field.Field
	ValueType   ValueType
}

MetricInfo contains the definition of a metric.

type MetricMetadata

type MetricMetadata struct {
	Units MetricDataUnits // the unit of recorded data for a given metric.
}

MetricMetadata contains user-provided metadata for a metric.

type Target

type Target interface {
	PopulateProto(d *pb.MetricsCollection)
	Hash() uint64
	Clone() Target
}

A Target knows how to put information about itself in a MetricsData message.

type ValueType

type ValueType int

ValueType is an enum for the type of a metric.

const (
	NonCumulativeIntType ValueType = iota
	CumulativeIntType
	NonCumulativeFloatType
	CumulativeFloatType
	StringType
	BoolType
	NonCumulativeDistributionType
	CumulativeDistributionType
)

Types of metric values.

func (ValueType) IsCumulative

func (v ValueType) IsCumulative() bool

IsCumulative returns true if this is a cumulative metric value type.

func (ValueType) String

func (v ValueType) String() string