1
0
mirror of https://github.com/open-telemetry/opentelemetry-go.git synced 2024-12-14 10:13:10 +02:00
opentelemetry-go/sdk/trace/span.go

829 lines
24 KiB
Go
Raw Normal View History

// Copyright The OpenTelemetry Authors
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package trace // import "go.opentelemetry.io/otel/sdk/trace"
import (
"context"
"fmt"
"reflect"
"runtime"
rt "runtime/trace"
"strings"
"sync"
"time"
"unicode/utf8"
"go.opentelemetry.io/otel/attribute"
"go.opentelemetry.io/otel/codes"
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
"go.opentelemetry.io/otel/sdk/instrumentation"
"go.opentelemetry.io/otel/sdk/internal"
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
"go.opentelemetry.io/otel/sdk/resource"
semconv "go.opentelemetry.io/otel/semconv/v1.21.0"
"go.opentelemetry.io/otel/trace"
)
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
// ReadOnlySpan allows reading information from the data structure underlying a
// trace.Span. It is used in places where reading information from a span is
// necessary but changing the span isn't necessary or allowed.
//
// Warning: methods may be added to this interface in minor releases.
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
type ReadOnlySpan interface {
// Name returns the name of the span.
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
Name() string
// SpanContext returns the unique SpanContext that identifies the span.
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
SpanContext() trace.SpanContext
// Parent returns the unique SpanContext that identifies the parent of the
// span if one exists. If the span has no parent the returned SpanContext
// will be invalid.
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
Parent() trace.SpanContext
// SpanKind returns the role the span plays in a Trace.
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
SpanKind() trace.SpanKind
// StartTime returns the time the span started recording.
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
StartTime() time.Time
// EndTime returns the time the span stopped recording. It will be zero if
// the span has not ended.
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
EndTime() time.Time
// Attributes returns the defining attributes of the span.
// The order of the returned attributes is not guaranteed to be stable across invocations.
Attributes() []attribute.KeyValue
// Links returns all the links the span has to other spans.
Links() []Link
// Events returns all the events that occurred within in the spans
// lifetime.
Events() []Event
// Status returns the spans status.
Status() Status
// InstrumentationScope returns information about the instrumentation
// scope that created the span.
InstrumentationScope() instrumentation.Scope
// InstrumentationLibrary returns information about the instrumentation
// library that created the span.
// Deprecated: please use InstrumentationScope instead.
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
InstrumentationLibrary() instrumentation.Library
// Resource returns information about the entity that produced the span.
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
Resource() *resource.Resource
// DroppedAttributes returns the number of attributes dropped by the span
// due to limits being reached.
DroppedAttributes() int
// DroppedLinks returns the number of links dropped by the span due to
// limits being reached.
DroppedLinks() int
// DroppedEvents returns the number of events dropped by the span due to
// limits being reached.
DroppedEvents() int
// ChildSpanCount returns the count of spans that consider the span a
// direct parent.
ChildSpanCount() int
// A private method to prevent users implementing the
// interface and so future additions to it will not
// violate compatibility.
private()
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
}
// ReadWriteSpan exposes the same methods as trace.Span and in addition allows
// reading information from the underlying data structure.
// This interface exposes the union of the methods of trace.Span (which is a
// "write-only" span) and ReadOnlySpan. New methods for writing or reading span
// information should be added under trace.Span or ReadOnlySpan, respectively.
//
// Warning: methods may be added to this interface in minor releases.
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
type ReadWriteSpan interface {
trace.Span
ReadOnlySpan
}
// recordingSpan is an implementation of the OpenTelemetry Span API
// representing the individual component of a trace that is sampled.
type recordingSpan struct {
// mu protects the contents of this span.
mu sync.Mutex
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
// parent holds the parent span of this span as a trace.SpanContext.
parent trace.SpanContext
// spanKind represents the kind of this span as a trace.SpanKind.
spanKind trace.SpanKind
// name is the name of this span.
name string
// startTime is the time at which this span was started.
startTime time.Time
// endTime is the time at which this span was ended. It contains the zero
// value of time.Time until the span is ended.
endTime time.Time
// status is the status of this span.
status Status
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
// childSpanCount holds the number of child spans created for this span.
childSpanCount int
// spanContext holds the SpanContext of this span.
spanContext trace.SpanContext
// attributes is a collection of user provided key/values. The collection
// is constrained by a configurable maximum held by the parent
// TracerProvider. When additional attributes are added after this maximum
// is reached these attributes the user is attempting to add are dropped.
// This dropped number of attributes is tracked and reported in the
// ReadOnlySpan exported when the span ends.
attributes []attribute.KeyValue
droppedAttributes int
// events are stored in FIFO queue capped by configured limit.
events evictedQueue
// links are stored in FIFO queue capped by configured limit.
links evictedQueue
// executionTracerTaskEnd ends the execution tracer span.
executionTracerTaskEnd func()
// tracer is the SDK tracer that created this span.
tracer *tracer
}
var _ ReadWriteSpan = (*recordingSpan)(nil)
var _ runtimeTracer = (*recordingSpan)(nil)
// SpanContext returns the SpanContext of this span.
func (s *recordingSpan) SpanContext() trace.SpanContext {
if s == nil {
return trace.SpanContext{}
}
return s.spanContext
}
// IsRecording returns if this span is being recorded. If this span has ended
// this will return false.
func (s *recordingSpan) IsRecording() bool {
if s == nil {
return false
}
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
s.mu.Lock()
defer s.mu.Unlock()
return s.endTime.IsZero()
}
// SetStatus sets the status of the Span in the form of a code and a
// description, overriding previous values set. The description is only
// included in the set status when the code is for an error. If this span is
// not being recorded than this method does nothing.
func (s *recordingSpan) SetStatus(code codes.Code, description string) {
if !s.IsRecording() {
return
}
s.mu.Lock()
defer s.mu.Unlock()
if s.status.Code > code {
return
}
status := Status{Code: code}
if code == codes.Error {
status.Description = description
}
s.status = status
}
// SetAttributes sets attributes of this span.
//
// If a key from attributes already exists the value associated with that key
// will be overwritten with the value contained in attributes.
//
// If this span is not being recorded than this method does nothing.
//
// If adding attributes to the span would exceed the maximum amount of
// attributes the span is configured to have, the last added attributes will
// be dropped.
func (s *recordingSpan) SetAttributes(attributes ...attribute.KeyValue) {
if !s.IsRecording() {
return
}
s.mu.Lock()
defer s.mu.Unlock()
limit := s.tracer.provider.spanLimits.AttributeCountLimit
if limit == 0 {
// No attributes allowed.
s.droppedAttributes += len(attributes)
return
}
// If adding these attributes could exceed the capacity of s perform a
// de-duplication and truncation while adding to avoid over allocation.
if limit > 0 && len(s.attributes)+len(attributes) > limit {
s.addOverCapAttrs(limit, attributes)
return
}
// Otherwise, add without deduplication. When attributes are read they
// will be deduplicated, optimizing the operation.
for _, a := range attributes {
if !a.Valid() {
// Drop all invalid attributes.
s.droppedAttributes++
continue
}
a = truncateAttr(s.tracer.provider.spanLimits.AttributeValueLengthLimit, a)
s.attributes = append(s.attributes, a)
}
}
// addOverCapAttrs adds the attributes attrs to the span s while
// de-duplicating the attributes of s and attrs and dropping attributes that
// exceed the limit.
//
// This method assumes s.mu.Lock is held by the caller.
//
// This method should only be called when there is a possibility that adding
// attrs to s will exceed the limit. Otherwise, attrs should be added to s
// without checking for duplicates and all retrieval methods of the attributes
// for s will de-duplicate as needed.
//
// This method assumes limit is a value > 0. The argument should be validated
// by the caller.
func (s *recordingSpan) addOverCapAttrs(limit int, attrs []attribute.KeyValue) {
// In order to not allocate more capacity to s.attributes than needed,
// prune and truncate this addition of attributes while adding.
// Do not set a capacity when creating this map. Benchmark testing has
// showed this to only add unused memory allocations in general use.
exists := make(map[attribute.Key]int)
s.dedupeAttrsFromRecord(&exists)
// Now that s.attributes is deduplicated, adding unique attributes up to
// the capacity of s will not over allocate s.attributes.
for _, a := range attrs {
if !a.Valid() {
// Drop all invalid attributes.
s.droppedAttributes++
continue
}
if idx, ok := exists[a.Key]; ok {
// Perform all updates before dropping, even when at capacity.
s.attributes[idx] = a
continue
}
if len(s.attributes) >= limit {
// Do not just drop all of the remaining attributes, make sure
// updates are checked and performed.
s.droppedAttributes++
} else {
a = truncateAttr(s.tracer.provider.spanLimits.AttributeValueLengthLimit, a)
s.attributes = append(s.attributes, a)
exists[a.Key] = len(s.attributes) - 1
}
}
}
// truncateAttr returns a truncated version of attr. Only string and string
// slice attribute values are truncated. String values are truncated to at
// most a length of limit. Each string slice value is truncated in this fashion
// (the slice length itself is unaffected).
//
// No truncation is performed for a negative limit.
func truncateAttr(limit int, attr attribute.KeyValue) attribute.KeyValue {
if limit < 0 {
return attr
}
switch attr.Value.Type() {
case attribute.STRING:
if v := attr.Value.AsString(); len(v) > limit {
return attr.Key.String(safeTruncate(v, limit))
}
case attribute.STRINGSLICE:
v := attr.Value.AsStringSlice()
for i := range v {
if len(v[i]) > limit {
v[i] = safeTruncate(v[i], limit)
}
}
return attr.Key.StringSlice(v)
}
return attr
}
// safeTruncate truncates the string and guarantees valid UTF-8 is returned.
func safeTruncate(input string, limit int) string {
if trunc, ok := safeTruncateValidUTF8(input, limit); ok {
return trunc
}
trunc, _ := safeTruncateValidUTF8(strings.ToValidUTF8(input, ""), limit)
return trunc
}
// safeTruncateValidUTF8 returns a copy of the input string safely truncated to
// limit. The truncation is ensured to occur at the bounds of complete UTF-8
// characters. If invalid encoding of UTF-8 is encountered, input is returned
// with false, otherwise, the truncated input will be returned with true.
func safeTruncateValidUTF8(input string, limit int) (string, bool) {
for cnt := 0; cnt <= limit; {
r, size := utf8.DecodeRuneInString(input[cnt:])
if r == utf8.RuneError {
return input, false
}
if cnt+size > limit {
return input[:cnt], true
}
cnt += size
}
return input, true
}
// End ends the span. This method does nothing if the span is already ended or
// is not being recorded.
//
// The only SpanOption currently supported is WithTimestamp which will set the
// end time for a Span's life-cycle.
//
// If this method is called while panicking an error event is added to the
// Span before ending it and the panic is continued.
func (s *recordingSpan) End(options ...trace.SpanEndOption) {
// Do not start by checking if the span is being recorded which requires
// acquiring a lock. Make a minimal check that the span is not nil.
if s == nil {
return
}
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
// Store the end time as soon as possible to avoid artificially increasing
// the span's duration in case some operation below takes a while.
et := internal.MonotonicEndTime(s.startTime)
// Do relative expensive check now that we have an end time and see if we
// need to do any more processing.
if !s.IsRecording() {
return
}
config := trace.NewSpanEndConfig(options...)
if recovered := recover(); recovered != nil {
// Record but don't stop the panic.
defer panic(recovered)
opts := []trace.EventOption{
trace.WithAttributes(
2023-02-07 23:42:47 +02:00
semconv.ExceptionType(typeStr(recovered)),
semconv.ExceptionMessage(fmt.Sprint(recovered)),
),
}
if config.StackTrace() {
opts = append(opts, trace.WithAttributes(
2023-02-07 23:42:47 +02:00
semconv.ExceptionStacktrace(recordStackTrace()),
))
}
s.addEvent(semconv.ExceptionEventName, opts...)
}
if s.executionTracerTaskEnd != nil {
s.executionTracerTaskEnd()
}
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
s.mu.Lock()
// Setting endTime to non-zero marks the span as ended and not recording.
if config.Timestamp().IsZero() {
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
s.endTime = et
} else {
s.endTime = config.Timestamp()
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
}
s.mu.Unlock()
sps := s.tracer.provider.getSpanProcessors()
if len(sps) == 0 {
return
}
snap := s.snapshot()
for _, sp := range sps {
sp.sp.OnEnd(snap)
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
}
}
// RecordError will record err as a span event for this span. An additional call to
// SetStatus is required if the Status of the Span should be set to Error, this method
// does not change the Span status. If this span is not being recorded or err is nil
// than this method does nothing.
func (s *recordingSpan) RecordError(err error, opts ...trace.EventOption) {
if s == nil || err == nil || !s.IsRecording() {
return
}
opts = append(opts, trace.WithAttributes(
2023-02-07 23:42:47 +02:00
semconv.ExceptionType(typeStr(err)),
semconv.ExceptionMessage(err.Error()),
))
c := trace.NewEventConfig(opts...)
if c.StackTrace() {
opts = append(opts, trace.WithAttributes(
2023-02-07 23:42:47 +02:00
semconv.ExceptionStacktrace(recordStackTrace()),
))
}
s.addEvent(semconv.ExceptionEventName, opts...)
}
func typeStr(i interface{}) string {
t := reflect.TypeOf(i)
if t.PkgPath() == "" && t.Name() == "" {
// Likely a builtin type.
return t.String()
}
return fmt.Sprintf("%s.%s", t.PkgPath(), t.Name())
}
func recordStackTrace() string {
stackTrace := make([]byte, 2048)
n := runtime.Stack(stackTrace, false)
return string(stackTrace[0:n])
}
// AddEvent adds an event with the provided name and options. If this span is
// not being recorded than this method does nothing.
func (s *recordingSpan) AddEvent(name string, o ...trace.EventOption) {
if !s.IsRecording() {
return
}
s.addEvent(name, o...)
}
func (s *recordingSpan) addEvent(name string, o ...trace.EventOption) {
c := trace.NewEventConfig(o...)
e := Event{Name: name, Attributes: c.Attributes(), Time: c.Timestamp()}
// Discard attributes over limit.
limit := s.tracer.provider.spanLimits.AttributePerEventCountLimit
if limit == 0 {
// Drop all attributes.
e.DroppedAttributeCount = len(e.Attributes)
e.Attributes = nil
} else if limit > 0 && len(e.Attributes) > limit {
// Drop over capacity.
e.DroppedAttributeCount = len(e.Attributes) - limit
e.Attributes = e.Attributes[:limit]
}
s.mu.Lock()
s.events.add(e)
s.mu.Unlock()
}
// SetName sets the name of this span. If this span is not being recorded than
// this method does nothing.
func (s *recordingSpan) SetName(name string) {
if !s.IsRecording() {
return
}
s.mu.Lock()
defer s.mu.Unlock()
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
s.name = name
}
// Name returns the name of this span.
func (s *recordingSpan) Name() string {
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
s.mu.Lock()
defer s.mu.Unlock()
return s.name
}
// Name returns the SpanContext of this span's parent span.
func (s *recordingSpan) Parent() trace.SpanContext {
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
s.mu.Lock()
defer s.mu.Unlock()
return s.parent
}
// SpanKind returns the SpanKind of this span.
func (s *recordingSpan) SpanKind() trace.SpanKind {
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
s.mu.Lock()
defer s.mu.Unlock()
return s.spanKind
}
// StartTime returns the time this span started.
func (s *recordingSpan) StartTime() time.Time {
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
s.mu.Lock()
defer s.mu.Unlock()
return s.startTime
}
// EndTime returns the time this span ended. For spans that have not yet
// ended, the returned value will be the zero value of time.Time.
func (s *recordingSpan) EndTime() time.Time {
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
s.mu.Lock()
defer s.mu.Unlock()
return s.endTime
}
// Attributes returns the attributes of this span.
//
// The order of the returned attributes is not guaranteed to be stable.
func (s *recordingSpan) Attributes() []attribute.KeyValue {
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
s.mu.Lock()
defer s.mu.Unlock()
s.dedupeAttrs()
return s.attributes
}
// dedupeAttrs deduplicates the attributes of s to fit capacity.
//
// This method assumes s.mu.Lock is held by the caller.
func (s *recordingSpan) dedupeAttrs() {
// Do not set a capacity when creating this map. Benchmark testing has
// showed this to only add unused memory allocations in general use.
exists := make(map[attribute.Key]int)
s.dedupeAttrsFromRecord(&exists)
}
// dedupeAttrsFromRecord deduplicates the attributes of s to fit capacity
// using record as the record of unique attribute keys to their index.
//
// This method assumes s.mu.Lock is held by the caller.
func (s *recordingSpan) dedupeAttrsFromRecord(record *map[attribute.Key]int) {
// Use the fact that slices share the same backing array.
unique := s.attributes[:0]
for _, a := range s.attributes {
if idx, ok := (*record)[a.Key]; ok {
unique[idx] = a
} else {
unique = append(unique, a)
(*record)[a.Key] = len(unique) - 1
}
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
}
// s.attributes have element types of attribute.KeyValue. These types are
// not pointers and they themselves do not contain pointer fields,
// therefore the duplicate values do not need to be zeroed for them to be
// garbage collected.
s.attributes = unique
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
}
// Links returns the links of this span.
func (s *recordingSpan) Links() []Link {
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
s.mu.Lock()
defer s.mu.Unlock()
if len(s.links.queue) == 0 {
return []Link{}
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
}
return s.interfaceArrayToLinksArray()
}
// Events returns the events of this span.
func (s *recordingSpan) Events() []Event {
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
s.mu.Lock()
defer s.mu.Unlock()
if len(s.events.queue) == 0 {
return []Event{}
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
}
return s.interfaceArrayToEventArray()
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
}
// Status returns the status of this span.
func (s *recordingSpan) Status() Status {
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
s.mu.Lock()
defer s.mu.Unlock()
return s.status
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
}
// InstrumentationScope returns the instrumentation.Scope associated with
// the Tracer that created this span.
func (s *recordingSpan) InstrumentationScope() instrumentation.Scope {
s.mu.Lock()
defer s.mu.Unlock()
return s.tracer.instrumentationScope
}
// InstrumentationLibrary returns the instrumentation.Library associated with
// the Tracer that created this span.
func (s *recordingSpan) InstrumentationLibrary() instrumentation.Library {
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
s.mu.Lock()
defer s.mu.Unlock()
return s.tracer.instrumentationScope
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
}
// Resource returns the Resource associated with the Tracer that created this
// span.
func (s *recordingSpan) Resource() *resource.Resource {
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
s.mu.Lock()
defer s.mu.Unlock()
return s.tracer.provider.resource
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
}
func (s *recordingSpan) addLink(link trace.Link) {
if !s.IsRecording() || !link.SpanContext.IsValid() {
return
}
l := Link{SpanContext: link.SpanContext, Attributes: link.Attributes}
// Discard attributes over limit.
limit := s.tracer.provider.spanLimits.AttributePerLinkCountLimit
if limit == 0 {
// Drop all attributes.
l.DroppedAttributeCount = len(l.Attributes)
l.Attributes = nil
} else if limit > 0 && len(l.Attributes) > limit {
l.DroppedAttributeCount = len(l.Attributes) - limit
l.Attributes = l.Attributes[:limit]
}
s.mu.Lock()
s.links.add(l)
s.mu.Unlock()
}
// DroppedAttributes returns the number of attributes dropped by the span
// due to limits being reached.
func (s *recordingSpan) DroppedAttributes() int {
s.mu.Lock()
defer s.mu.Unlock()
return s.droppedAttributes
}
// DroppedLinks returns the number of links dropped by the span due to limits
// being reached.
func (s *recordingSpan) DroppedLinks() int {
s.mu.Lock()
defer s.mu.Unlock()
return s.links.droppedCount
}
// DroppedEvents returns the number of events dropped by the span due to
// limits being reached.
func (s *recordingSpan) DroppedEvents() int {
s.mu.Lock()
defer s.mu.Unlock()
return s.events.droppedCount
}
// ChildSpanCount returns the count of spans that consider the span a
// direct parent.
func (s *recordingSpan) ChildSpanCount() int {
s.mu.Lock()
defer s.mu.Unlock()
return s.childSpanCount
}
// TracerProvider returns a trace.TracerProvider that can be used to generate
// additional Spans on the same telemetry pipeline as the current Span.
func (s *recordingSpan) TracerProvider() trace.TracerProvider {
return s.tracer.provider
}
// snapshot creates a read-only copy of the current state of the span.
func (s *recordingSpan) snapshot() ReadOnlySpan {
var sd snapshot
s.mu.Lock()
defer s.mu.Unlock()
sd.endTime = s.endTime
sd.instrumentationScope = s.tracer.instrumentationScope
sd.name = s.name
sd.parent = s.parent
sd.resource = s.tracer.provider.resource
sd.spanContext = s.spanContext
sd.spanKind = s.spanKind
sd.startTime = s.startTime
sd.status = s.status
sd.childSpanCount = s.childSpanCount
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
if len(s.attributes) > 0 {
s.dedupeAttrs()
sd.attributes = s.attributes
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
}
sd.droppedAttributeCount = s.droppedAttributes
if len(s.events.queue) > 0 {
sd.events = s.interfaceArrayToEventArray()
sd.droppedEventCount = s.events.droppedCount
}
if len(s.links.queue) > 0 {
sd.links = s.interfaceArrayToLinksArray()
sd.droppedLinkCount = s.links.droppedCount
}
return &sd
}
func (s *recordingSpan) interfaceArrayToLinksArray() []Link {
linkArr := make([]Link, 0)
for _, value := range s.links.queue {
linkArr = append(linkArr, value.(Link))
}
return linkArr
}
func (s *recordingSpan) interfaceArrayToEventArray() []Event {
eventArr := make([]Event, 0)
for _, value := range s.events.queue {
eventArr = append(eventArr, value.(Event))
}
return eventArr
}
func (s *recordingSpan) addChild() {
if !s.IsRecording() {
return
}
s.mu.Lock()
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
s.childSpanCount++
s.mu.Unlock()
}
func (*recordingSpan) private() {}
// runtimeTrace starts a "runtime/trace".Task for the span and returns a
// context containing the task.
func (s *recordingSpan) runtimeTrace(ctx context.Context) context.Context {
if !rt.IsEnabled() {
// Avoid additional overhead if runtime/trace is not enabled.
return ctx
}
nctx, task := rt.NewTask(ctx, s.name)
s.mu.Lock()
s.executionTracerTaskEnd = task.End
s.mu.Unlock()
return nctx
}
// nonRecordingSpan is a minimal implementation of the OpenTelemetry Span API
// that wraps a SpanContext. It performs no operations other than to return
// the wrapped SpanContext or TracerProvider that created it.
type nonRecordingSpan struct {
// tracer is the SDK tracer that created this span.
tracer *tracer
sc trace.SpanContext
}
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
var _ trace.Span = nonRecordingSpan{}
// SpanContext returns the wrapped SpanContext.
func (s nonRecordingSpan) SpanContext() trace.SpanContext { return s.sc }
// IsRecording always returns false.
func (nonRecordingSpan) IsRecording() bool { return false }
// SetStatus does nothing.
func (nonRecordingSpan) SetStatus(codes.Code, string) {}
Add RO/RW span interfaces (#1360) * Store span data directly in the span - Nesting only some of a span's data in a `data` field (with the rest of the data living direclty in the `span` struct) is confusing. - export.SpanData is meant to be an immutable *snapshot* of a span, not the "authoritative" state of the span. - Refactor attributesMap.toSpanData into toKeyValue and make it return a []label.KeyValue which is clearer than modifying a struct passed to the function. - Read droppedCount from the attributesMap as a separate operation instead of setting it from within attributesMap.toSpanData. - Set a span's end time in the span itself rather than in the SpanData to allow reading the span's end time after a span has ended. - Set a span's end time as soon as possible within span.End so that we don't influence the span's end time with operations such as fetching span processors and generating span data. - Remove error handling for uninitialized spans. This check seems to be necessary only because we used to have an *export.SpanData field which could be nil. Now that we no longer have this field I think we can safely remove the check. The error isn't used anywhere else so remove it, too. * Store parent as trace.SpanContext The spec requires that the parent field of a Span be a Span, a SpanContext or null. Rather than extracting the parent's span ID from the trace.SpanContext which we get from the tracer, store the trace.SpanContext as is and explicitly extract the parent's span ID where necessary. * Add ReadOnlySpan interface Use this interface instead of export.SpanData in places where reading information from a span is necessary. Use export.SpanData only when exporting spans. * Add ReadWriteSpan interface Use this interface instead of export.SpanData in places where it is necessary to read information from a span and write to it at the same time. * Rename export.SpanData to SpanSnapshot SpanSnapshot represents the nature of this type as well as its intended use more accurately. Clarify the purpose of SpanSnapshot in the docs and emphasize what should and should not be done with it. * Rephrase attributesMap doc comment "refreshes" is wrong for plural ("updates"). * Refactor span.End() - Improve accuracy of span duration. Record span end time ASAP. We want to measure a user operation as accurately as possible, which means we want to mark the end time of a span as soon as possible after span.End() is called. Any operations we do inside span.End() before storing the end time affect the total duration of the span, and although these operations are rather fast at the moment they still seem to affect the duration of the span by "artificially" adding time between the start and end timestamps. This is relevant only in cases where the end time isn't explicitly specified. - Remove redundant idempotence check. Now that IsRecording() is based on the value of span.endTime, IsRecording() will always return false after span.End() had been called because span.endTime won't be zero. This means we no longer need span.endOnce. - Improve TestEndSpanTwice so that it also ensures subsequent calls to span.End() don't modify the span's end time. * Update changelog Co-authored-by: Tyler Yahn <codingalias@gmail.com> Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
2020-12-11 07:15:44 +02:00
// SetError does nothing.
func (nonRecordingSpan) SetError(bool) {}
// SetAttributes does nothing.
func (nonRecordingSpan) SetAttributes(...attribute.KeyValue) {}
// End does nothing.
func (nonRecordingSpan) End(...trace.SpanEndOption) {}
// RecordError does nothing.
func (nonRecordingSpan) RecordError(error, ...trace.EventOption) {}
// AddEvent does nothing.
func (nonRecordingSpan) AddEvent(string, ...trace.EventOption) {}
// SetName does nothing.
func (nonRecordingSpan) SetName(string) {}
// TracerProvider returns the trace.TracerProvider that provided the Tracer
// that created this span.
func (s nonRecordingSpan) TracerProvider() trace.TracerProvider { return s.tracer.provider }
func isRecording(s SamplingResult) bool {
return s.Decision == RecordOnly || s.Decision == RecordAndSample
}
func isSampled(s SamplingResult) bool {
return s.Decision == RecordAndSample
}
// Status is the classified state of a Span.
type Status struct {
// Code is an identifier of a Spans state classification.
Code codes.Code
// Description is a user hint about why that status was set. It is only
// applicable when Code is Error.
Description string
}