Support distributed trace propagation across SQS

[esmiralha]

Package Name

aws-sdk

Package Version(s)

No response

Describe the feature you'd like

SQS is a distributed queue by amazon. It allows metadata with message attributes so implementation should be fairly straightforward.

Is your feature request related to a problem?

No response

Describe alternatives you've considered

No response

Additional context

No response

[darccio]

Hi @esmiralha, do you miss any feature around SQS tracing? It's already supported in both versions for aws-sdk-go:

• v1
• v2

v1 works by wrapping the session (as you can see as done by Orchestrion here), and v2 achieves the same by wrapping the config (as you can see done by Orchestrion here).

[rarguelloF]

Hello! 👋

Yes, distributed context propagation should be supported for SQS.
It is strongly recommended to use the AWS SDK v2 for Go (and the corresponding v2 integration in dd-trace-go), as v1 was deprecated by AWS and therefore our integration is deprecated too.

Please let us know if you find any issues while using the integration!

[nhulston]

Hi @esmiralha, you will also need to add this line of code: awstrace.AppendMiddleware(&cfg)

Here is a simple working example:

main.go:

package main

import (
	"context"
	"log"

	"github.com/aws/aws-sdk-go-v2/aws"
	"github.com/aws/aws-sdk-go-v2/config"
	"github.com/aws/aws-sdk-go-v2/service/sqs"
	awstrace "github.com/DataDog/dd-trace-go/contrib/aws/aws-sdk-go-v2/v2/aws"
	"github.com/DataDog/dd-trace-go/v2/ddtrace/tracer"
)

func main() {
	// Initialize Datadog tracer
	tracer.Start(tracer.WithService("go-sqs-test"))
	defer tracer.Stop()

	// Create AWS config
	ctx := context.Background()
	cfg, err := config.LoadDefaultConfig(ctx, config.WithRegion("<REGION>"))
	if err != nil {
		log.Fatal(err)
	}

	// Add Datadog tracing middleware
	awstrace.AppendMiddleware(&cfg)

	// Create SQS client
	sqsClient := sqs.NewFromConfig(cfg)

	// Send message
	result, err := sqsClient.SendMessage(ctx, &sqs.SendMessageInput{
		QueueUrl:    aws.String("https://sqs.<REGION>.amazonaws.com/<ACCONT_ID>/<QUEUE_NAME>"),
		MessageBody: aws.String("foobar"),
	})

	if err != nil {
		log.Fatal(err)
	}

	log.Printf("Message sent! MessageId: %s", *result.MessageId)
}

go.mod:

module go-sqs-test

go 1.24.0

require (
	github.com/DataDog/dd-trace-go/contrib/aws/aws-sdk-go-v2/v2 v2.4.0
	github.com/DataDog/dd-trace-go/v2 v2.4.0
	github.com/aws/aws-sdk-go-v2 v1.40.0
	github.com/aws/aws-sdk-go-v2/config v1.32.2
	github.com/aws/aws-sdk-go-v2/service/sqs v1.42.17
)

And you can see the Datadog trace context is injected into the message attributes:

Please let us know if you run into more issues. Thanks

[esmiralha]

thanks for the prompt reply!
Our code looks exactly like that and yet the traces are not propagating. But given that it is expected to propagate, then I probably need to debug the issue on my side.
Thanks again.

[nhulston]

What is the queue triggering? As far as I know, only the Lambda extension has logic for extracting and reparenting the trace context.

If you are triggering something else (not Lambda), you may have to write the trace extraction & reparenting logic yourself, or submit a feature request.

[esmiralha]

It's not a Lambda, but rather ECS services which are polling SQS and handling batches of up to 10 messages.
I guess you got to the bottom of the question. I need to manually extract & reparent.
My remaining doubt is whether I have to also manually inject the trace into the message when publishing or if the library will take care of that.
Thanks!