Skip to content

OpenTelemetry: Specify span kind and don't set error type for benign exception #2048

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 3 commits into
base: master
Choose a base branch
from
Open

OpenTelemetry: Specify span kind and don't set error type for benign exception #2048

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
c2bb055
Add specifying SpanKind for Server and Client
yuandrew Sep 10, 2025
b637997
Don't set error for benign exceptions
yuandrew Sep 10, 2025
72a76d4
don't set error status for benign exception errors
yuandrew Sep 10, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
21 changes: 19 additions & 2 deletions contrib/opentelemetry/tracing_interceptor.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,6 +3,7 @@ package opentelemetry

import (
"context"
"errors"
"fmt"

"go.opentelemetry.io/otel"
Expand All @@ -14,6 +15,7 @@ import (

"go.temporal.io/sdk/interceptor"
"go.temporal.io/sdk/log"
"go.temporal.io/sdk/temporal"
)

// DefaultTextMapPropagator is the default OpenTelemetry TextMapPropagator used
Expand Down Expand Up @@ -196,8 +198,13 @@ func (t *tracer) StartSpan(opts *interceptor.TracerStartSpanOptions) (intercepto
}
}

spanKind := trace.SpanKindServer
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Does this line up roughly with what other SDKs like dotnet or python do?

Copy link
Member

@cretz cretz Sep 11, 2025 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think so (though I think I just spotted a bug in Python for outbound signal child workflow)

if opts.Outbound {
spanKind = trace.SpanKindClient
}

// Create span
span := t.options.SpanStarter(ctx, t.options.Tracer, opts.Operation+":"+opts.Name, trace.WithTimestamp(opts.Time))
span := t.options.SpanStarter(ctx, t.options.Tracer, opts.Operation+":"+opts.Name, trace.WithTimestamp(opts.Time), trace.WithSpanKind(spanKind))

// Set tags
if len(opts.Tags) > 0 {
Expand Down Expand Up @@ -241,12 +248,22 @@ type tracerSpan struct {
}

func (t *tracerSpan) Finish(opts *interceptor.TracerFinishSpanOptions) {
if opts.Error != nil {
if opts.Error != nil && !isBenignApplicationError(opts.Error) {
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think our implementation may not have been doing the right thing here because we are not calling .RecordError, which I think we should always do on error regardless of whether benign (but status as error only for benign like you have here). I don't know if we consider now starting to record errors as a breaking/dangerous change, but probably not.

t.SetStatus(codes.Error, opts.Error.Error())
}
t.End()
}

func isBenignApplicationError(err error) bool {
var appErr *temporal.ApplicationError
if temporal.IsApplicationError(err) {
if errors.As(err, &appErr) {
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Per #1925 (comment), we should only look at the top-level error to determine whether benign. Mimic the isBenignApplicationError already in internal/error.go.

return appErr.Category() == temporal.ApplicationErrorCategoryBenign
}
}
return false
}

type textMapCarrier map[string]string

func (t textMapCarrier) Get(key string) string { return t[key] }
Expand Down
106 changes: 106 additions & 0 deletions contrib/opentelemetry/tracing_interceptor_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,15 +2,19 @@ package opentelemetry_test

import (
"testing"
"time"

"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
"go.opentelemetry.io/otel/codes"
sdktrace "go.opentelemetry.io/otel/sdk/trace"
"go.opentelemetry.io/otel/sdk/trace/tracetest"
"go.opentelemetry.io/otel/trace"

"go.temporal.io/sdk/contrib/opentelemetry"
"go.temporal.io/sdk/interceptor"
"go.temporal.io/sdk/internal/interceptortest"
"go.temporal.io/sdk/temporal"
)

func TestSpanPropagation(t *testing.T) {
Expand Down Expand Up @@ -42,3 +46,105 @@ func spanChildren(spans []sdktrace.ReadOnlySpan, parentID trace.SpanID) (ret []*
}
return
}

func TestSpanKind(t *testing.T) {
tests := []struct {
operation string
outbound bool
expectedKind trace.SpanKind
}{
{
operation: "StartWorkflow",
outbound: true,
expectedKind: trace.SpanKindClient,
},
{
operation: "RunWorkflow",
outbound: false,
expectedKind: trace.SpanKindServer,
},
}

for _, tt := range tests {
t.Run(tt.operation, func(t *testing.T) {
rec := tracetest.NewSpanRecorder()
tracer, err := opentelemetry.NewTracer(opentelemetry.TracerOptions{
Tracer: sdktrace.NewTracerProvider(sdktrace.WithSpanProcessor(rec)).Tracer(""),
})
require.NoError(t, err)

span, err := tracer.StartSpan(&interceptor.TracerStartSpanOptions{
Operation: tt.operation,
Name: "test-span",
Outbound: tt.outbound,
})
require.NoError(t, err)

span.Finish(&interceptor.TracerFinishSpanOptions{})

spans := rec.Ended()
require.Equal(t, len(spans), 1)

foundSpan := spans[0]
assert.Equal(t, tt.expectedKind, foundSpan.SpanKind(),
"Expected span kind %v but got %v for operation %s (outbound=%v)",
tt.expectedKind, foundSpan.SpanKind(), tt.operation, tt.outbound)
})
}
}

func TestBenignErrorSpanStatus(t *testing.T) {
tests := []struct {
name string
err error
expectError bool
expectStatus codes.Code
}{
{
name: "benign application error should not set error status",
err: temporal.NewApplicationErrorWithOptions("benign error", "TestType", temporal.ApplicationErrorOptions{Category: temporal.ApplicationErrorCategoryBenign}),
expectError: false,
expectStatus: codes.Unset,
},
{
name: "regular application error should set error status",
err: temporal.NewApplicationError("regular error", "TestType"),
expectError: true,
expectStatus: codes.Error,
},
}

for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
rec := tracetest.NewSpanRecorder()
tracer, err := opentelemetry.NewTracer(opentelemetry.TracerOptions{
Tracer: sdktrace.NewTracerProvider(sdktrace.WithSpanProcessor(rec)).Tracer(""),
})
require.NoError(t, err)

span, err := tracer.StartSpan(&interceptor.TracerStartSpanOptions{
Operation: "TestOperation",
Name: "TestSpan",
Time: time.Now(),
})
require.NoError(t, err)

span.Finish(&interceptor.TracerFinishSpanOptions{
Error: tt.err,
})

// Check recorded spans
spans := rec.Ended()
require.Len(t, spans, 1)

recordedSpan := spans[0]
assert.Equal(t, tt.expectStatus, recordedSpan.Status().Code)

if tt.expectError {
assert.NotEmpty(t, recordedSpan.Status().Description)
} else {
assert.Empty(t, recordedSpan.Status().Description)
}
})
}
}
13 changes: 12 additions & 1 deletion contrib/opentracing/interceptor.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,6 +5,7 @@ import (
"context"
"errors"
"fmt"
"go.temporal.io/sdk/temporal"

"github.com/opentracing/opentracing-go"

Expand Down Expand Up @@ -164,9 +165,19 @@ type tracerSpanRef struct{ opentracing.SpanContext }
type tracerSpan struct{ opentracing.Span }

func (t *tracerSpan) Finish(opts *interceptor.TracerFinishSpanOptions) {
if opts.Error != nil {
if opts.Error != nil && !isBenignApplicationError(opts.Error) {
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Open tracing is deprecated, not sure we should bother fixing it here/ do a release to publish this

// Standard tag that can be bridged to OpenTelemetry
t.SetTag("error", "true")
}
t.Span.Finish()
}

func isBenignApplicationError(err error) bool {
var appErr *temporal.ApplicationError
if temporal.IsApplicationError(err) {
if errors.As(err, &appErr) {
return appErr.Category() == temporal.ApplicationErrorCategoryBenign
}
}
return false
}
Loading
Loading