telemetry

Documentation

Index

• func Cleanup() error
• func GetTraceClient() (client otlptrace.Client, err error)
• func Must[T any](val T, err error) T
• func RecordErrorOnSpan(span trace.Span) func(error) error
• func RecordErrorOnSpanOneChannel[T any](span trace.Span) func(<-chan T, error) (<-chan T, error)
• func RecordErrorOnSpanReadCloserAndClose(span trace.Span) func(io.ReadCloser, error) (io.ReadCloser, error)
• func RecordErrorOnSpanReadCloserTwoAndClose[T any](span trace.Span) func(io.ReadCloser, T, error) (io.ReadCloser, T, error)
• func RecordErrorOnSpanThree[T any, S any](span trace.Span) func(T, S, error) (T, S, error)
• func RecordErrorOnSpanTwo[T any](span trace.Span) func(T, error) (T, error)
• func RecordErrorOnSpanTwoChannels[T any](span trace.Span) func(<-chan T, <-chan error) (<-chan T, <-chan error)
• func SetupFromEnvs()
• func Timer(ctx context.Context, durationRecorder metric.Int64Histogram, ...) func() time.Duration
• type Counter
  • func NewCounter(meter metric.Meter, name string, description string) (*Counter, error)
  • func (c *Counter) Add(ctx context.Context, num int64, attrs ...attribute.KeyValue)
  • func (c *Counter) Inc(ctx context.Context, attrs ...attribute.KeyValue)
• type Gauge
  • func NewGauge(meter metric.Meter, name string, description string) (*Gauge, error)
  • func (g *Gauge) Dec(ctx context.Context, attrs ...attribute.KeyValue)
  • func (g *Gauge) Inc(ctx context.Context, attrs ...attribute.KeyValue)
• type MetricRecorder
  • func NewMetricRecorder() *MetricRecorder
  • func (t *MetricRecorder) RecordFault(ctx context.Context, counter metric.Int64Counter, options ...metric.AddOption)
  • func (t *MetricRecorder) RecordFaultAndLatency(ctx context.Context, counter metric.Int64Counter, ...)
  • func (t *MetricRecorder) RecordLatency(ctx context.Context, histogram metric.Int64Histogram, ...)
  • func (t *MetricRecorder) RecordTotalLatency(ctx context.Context, histogram metric.Int64Histogram, ...)
  • func (t *MetricRecorder) Success()

Functions

func Cleanup

func Cleanup() error

Cleanup flushes the remaining traces and metrics in memory to the exporter and releases any telemetry resources.

func GetTraceClient

func GetTraceClient() (client otlptrace.Client, err error)

func Must

func Must[T any](val T, err error) T

func RecordErrorOnSpan

func RecordErrorOnSpan(span trace.Span) func(error) error

RecordErrorOnSpan records the error returned by the function in the given span. Should be used like `return RecordErrorOnSpan(span)(c.client.NetworkRemove(ctx, networkID))`

func RecordErrorOnSpanOneChannel

func RecordErrorOnSpanOneChannel[T any](span trace.Span) func(<-chan T, error) (<-chan T, error)

RecordErrorOnSpanOneChannel is similar to RecordErrorOnSpanTwo but one parameter is a channel.

func RecordErrorOnSpanReadCloserAndClose

func RecordErrorOnSpanReadCloserAndClose(span trace.Span) func(io.ReadCloser, error) (io.ReadCloser, error)

RecordErrorOnSpanReadCloserAndClose is similar to RecordError but takes an additional parameter that is an io.ReadCloser. The function will end the given span when the io.ReadCloser is closed. The caller is expected to _not_ call end on the span.

func RecordErrorOnSpanReadCloserTwoAndClose

func RecordErrorOnSpanReadCloserTwoAndClose[T any](span trace.Span) func(io.ReadCloser, T, error) (io.ReadCloser, T, error)

RecordErrorOnSpanReadCloserTwoAndClose is similar to RecordErrorOnSpanReadCloserAndClose but takes an additional parameter.

func RecordErrorOnSpanThree

func RecordErrorOnSpanThree[T any, S any](span trace.Span) func(T, S, error) (T, S, error)

RecordErrorOnSpanThree is similar to RecordError but the function being called takes an additional parameter.

func RecordErrorOnSpanTwo

func RecordErrorOnSpanTwo[T any](span trace.Span) func(T, error) (T, error)

RecordErrorOnSpanTwo is similar to RecordErrorOnSpan but the function being called takes an additional parameter.

func RecordErrorOnSpanTwoChannels

func RecordErrorOnSpanTwoChannels[T any](span trace.Span) func(<-chan T, <-chan error) (<-chan T, <-chan error)

RecordErrorOnSpanTwoChannels is similar to RecordErrorOnSpanTwo but both parameters are channels.

func SetupFromEnvs

func SetupFromEnvs()

func Timer

func Timer(
	ctx context.Context,
	durationRecorder metric.Int64Histogram,
	attrs ...attribute.KeyValue,
) func() time.Duration

Timer is a function to record a duration. Calling it starts the timer, calling the returned function will record the duration.

Types

type Counter

type Counter struct {
	// contains filtered or unexported fields
}

Counter is a synchronous Instrument which supports non-negative increments Example uses for Counter: - count the number of bytes received - count the number of requests completed - count the number of accounts created - count the number of checkpoints run - count the number of HTTP 5xx errors

func NewCounter

func NewCounter(meter metric.Meter, name string, description string) (*Counter, error)

func (*Counter) Add

func (c *Counter) Add(ctx context.Context, num int64, attrs ...attribute.KeyValue)

func (*Counter) Inc

func (c *Counter) Inc(ctx context.Context, attrs ...attribute.KeyValue)

type Gauge

type Gauge struct {
	// contains filtered or unexported fields
}

Gauge is a synchronous Instrument which supports increments and decrements. Note: if the value is monotonically increasing, use Counter instead. Example uses for Gauge: - the number of active requests - the number of items in a queue

func NewGauge

func NewGauge(meter metric.Meter, name string, description string) (*Gauge, error)

func (*Gauge) Dec

func (g *Gauge) Dec(ctx context.Context, attrs ...attribute.KeyValue)

func (*Gauge) Inc

func (g *Gauge) Inc(ctx context.Context, attrs ...attribute.KeyValue)

type MetricRecorder

type MetricRecorder struct {
	// contains filtered or unexported fields
}

MetricRecorder is a helper for recording metrics. Specifically, it provides methods to record latency and faults.

Example usage:

recorder := NewMetricRecorder()
defer recorder.RecordFault(ctx, faultCounter) // Records fault if success was not reported
defer recorder.RecordTotalLatency(ctx, totalLatency) // Records total latency since start

// Alternatively, you can use the combined method to record fault and total latency
// defer recorder.RecordFaultAndLatency(ctx, faultCounter, totalLatency)

msg := queue.Receive()
recorder.RecordLatency(ctx, dequeueLatency) // Records dequeue latency since start

process(msg)
recorder.RecordLatency(ctx, processLatency) // Records process latency since queue.Receive()

recorder.Success()

func NewMetricRecorder

func NewMetricRecorder() *MetricRecorder

func (*MetricRecorder) RecordFault

func (t *MetricRecorder) RecordFault(ctx context.Context, counter metric.Int64Counter, options ...metric.AddOption)

RecordFault records a fault as 1 if success was not reported, 0 otherwise.

func (*MetricRecorder) RecordFaultAndLatency

func (t *MetricRecorder) RecordFaultAndLatency(
	ctx context.Context, counter metric.Int64Counter, histogram metric.Int64Histogram, options ...metric.MeasurementOption)

RecordFaultAndLatency records a fault and total latency.

func (*MetricRecorder) RecordLatency

func (t *MetricRecorder) RecordLatency(ctx context.Context, histogram metric.Int64Histogram, options ...metric.RecordOption)

RecordLatency records the latency since the last operation. If this the first operation, it records the latency since the start.

func (*MetricRecorder) RecordTotalLatency

func (t *MetricRecorder) RecordTotalLatency(ctx context.Context, histogram metric.Int64Histogram, options ...metric.RecordOption)

RecordTotalLatency records the latency since the start.

func (*MetricRecorder) Success

func (t *MetricRecorder) Success()

Success records that the operation was successful. It doesn't record anything, but it is used to determine whether a fault should be recorded.