Add helper methods for manual spans in mutiny pipelines

[arn-ivu]

Fix #44411

In a traditional blocking and synchronous framework the opentelemetry context is attached to ThreadLocal. In reactive programming, where multiple processings share the same event-loop thread, one has to use Vert.x duplicated contexts instead (see https://quarkus.io/guides/duplicated-context). wrapWithSpan ensures that the pipeline is executed on a duplicated context (If the current context already is duplicated, it will stay the same. Therefore, nested calls to wrapWithSpan will all run on the same vert.x context).

Another difficulty of mutiny pipelines is, that usually only one parameter flows through the pipeline. In oder to end spans and to close the current scope at the end of the pipeline, span and scope must be stored in the mutiny-context. They are stored in a stack datastructure so that multiple nested spans/scopes are closed in the correct order.

inspired by Jan Peremsky (https://github.com/jan-peremsky/quarkus-reactive-otel/blob/c74043d388ec4df155f466f1d6938931c3389b70/src/main/java/com/fng/ewallet/pex/Tracer.java) and edeandrea (https://github.com/quarkusio/quarkus-super-heroes/blob/main/event-statistics/src/main/java/io/quarkus/sample/superheroes/statistics/listener/SuperStats.java)

[quarkus-bot]

Thanks for your pull request!

Your pull request does not follow our editorial rules. Could you have a look?

• title should not end up with dot

_{This message is automatically generated by a bot.}

[quarkus-bot]

/cc @brunobat (opentelemetry), @radcortez (opentelemetry)

[brunobat]

Sorry, Please use the injected OpenTelemetry instance. No need for any of this setup.
You can also inject the tracer and span.

[arn-ivu]

I think to do that, I would have to use QuarkusUnitTest-Extension, right?
Unfortunately QuarkusUnitTest together with ParameterizedTests does not seem to work. I think I would then do only one of the tests on all three contexts, and the rest of the tests on root-context only.
What do you think @brunobat? Or do I overlook a simpler solution?

[arn-ivu]

I have pushed an update with the altered tests.

[brunobat]

There are a few examples that work with parameterised tests:

quarkus/extensions/funqy/funqy-http/deployment/src/test/java/io/quarkus/funqy/test/SimpleTest.java

Line 17 in 01bee69

public class SimpleTest {

and

quarkus/extensions/funqy/funqy-knative-events/deployment/src/test/java/io/quarkus/funqy/test/ExposedCloudEventTest.java

Line 21 in 5180a5a

public class ExposedCloudEventTest {

[github-actions]

🎊 PR Preview f010b20 has been successfully built and deployed to https://quarkus-pr-main-45478-preview.surge.sh/version/main/guides/

• Images of blog posts older than 3 months are not available.
• Newsletters older than 3 months are not available.

[quarkus-bot]

Status for workflow Quarkus CI

This is the status report for running Quarkus CI on commit dc05142.

Failing Jobs

Status	Name	Step	Failures	Logs	Raw logs	Build scan
✖	Initial JDK 17 Build	Build	Failures	Logs	Raw logs	🔍

You can consult the Develocity build scans.

Failures

⚙️ Initial JDK 17 Build #

- Failing: extensions/opentelemetry/deployment 
! Skipped: devtools/bom-descriptor-json docs extensions/liquibase-mongodb/deployment and 62 more

📦 extensions/opentelemetry/deployment

✖ Failed to execute goal net.revelc.code:impsort-maven-plugin:1.12.0:check (check-imports) on project quarkus-opentelemetry-deployment: Imports are not sorted in /home/runner/work/quarkus/quarkus/extensions/opentelemetry/deployment/src/test/java/io/quarkus/opentelemetry/deployment/traces/MutinyTracingHelperTest.java

[quarkus-bot]

Status for workflow Quarkus Documentation CI

This is the status report for running Quarkus Documentation CI on commit dc05142.

✅ The latest workflow run for the pull request has completed successfully.

It should be safe to merge provided you have a look at the other checks in the summary.

[ozangunalp]

I need to read the motivation behind the Tracer method linked to the issue itself. For me, this needs to be equivalent to the WithSpanInterceptor, ie. an extracted method returning Uni and annotated with @WithSpan should have the exact same effect.

A helper method like this doesn't need to be redundant, but I don't see why it wouldn't be implemented the same way.

As for what has been done in SuperStats, this has to do with handling message traces. I need to set up an environment myself to compare produced traces, but I am pretty sure such code is no longer needed for message traces.

[arn-ivu]

@ozangunalp My usecase looks roughly like this

private void handleMessages() {
        consumer.getMessages() //from our own rabbitMQ-Lib. Will issue item whenever a message arrives
                .onItem()
                .transformToUniAndConcatenate(message -> wrapWithSpan(tracer, 
                                                //this is how our lib provides the remote context
						message.getTracingData().map(MessageTracingData::getContext), 
                                               "handleIncomingMessage", 
						Uni.createFrom()
                                .item(message)
                                .map(Message::getPayload)
                                .invoke(this::handlePayload) //actual message processing
                )
                .subscribe().with(...);
    }

it might be possible to solve it like this

private void handleMessages() {
        consumer.getMessages() //from our own rabbitMQ-Lib. Will issue item whenever a message arrives
                .onItem().invoke(message -> makeContextCurrent(message.getTracingData().map(MessageTracingData::getContext)))
                .onItem().call(this::handleIncomingMessage)
                .subscribe().with(...)
    }
    
@WithSpan
public Uni<Void> handleMessage(Message message){
	Uni.createFrom()
    	.item(message)
        .map(Message::getPayload)
        .invoke(this::handlePayload) //actual message processing
        .replaceWithVoid()
}

however, I have 2 problems with that

1. I would have to expose the extracted Method with public (otherwise I get a warning that @WithSpan won't do anything)
2. to make the provided context current (in order for WithSpan to set the parent correctly) I would have to make sure to run on a duplicate context and I think I would have to close the context later as well. So essentially the same problems I encounter when creating the span manually.