Kotlin grpc context validateGeneration error

[KarinaGanieva-sl]

Describe the bug

My vertx kotlin application runs runBlocking code and makes a great number of grpc calls to the server. I'm using javaagent (version 2.4.0) for auto-instrumentation on my client.
Making grpc calls with kotlin grpc stub and collecting flow leads to this error:
May 20, 2024 8:21:17 PM io.grpc.Context validateGeneration SEVERE: Context ancestry chain length is abnormally long. This suggests an error in application code. Length exceeded: 1000 java.lang.Exception at io.grpc.Context.validateGeneration(Context.java:1110) at io.grpc.Context.<init>(Context.java:191) at io.grpc.Context.fork(Context.java:404) at io.opentelemetry.javaagent.shaded.instrumentation.grpc.v1_6.internal.ContextStorageBridge.current(ContextStorageBridge.java:138) at io.grpc.Context.current(Context.java:172) at opentelemetry.GrpcClientKotlinStub$Companion$main$1.invokeSuspend(GrpcClientKotlinStub.kt:53) at kotlin.coroutines.jvm.internal.BaseContinuationImpl.resumeWith(ContinuationImpl.kt:33) at kotlinx.coroutines.internal.ScopeCoroutine.afterResume(Scopes.kt:28) at kotlinx.coroutines.AbstractCoroutine.resumeWith(AbstractCoroutine.kt:99) at kotlin.coroutines.jvm.internal.BaseContinuationImpl.resumeWith(ContinuationImpl.kt:46) at kotlinx.coroutines.DispatchedTask.run(DispatchedTask.kt:104) at io.reactivex.Scheduler$DisposeTask.run(Scheduler.java:608) at io.vertx.reactivex.ContextScheduler$ContextWorker$TimedAction.run(ContextScheduler.java:190) at io.vertx.reactivex.ContextScheduler$ContextWorker$TimedAction.lambda$execute$1(ContextScheduler.java:175) at io.vertx.core.impl.ContextImpl.lambda$executeBlocking$1(ContextImpl.java:191) at io.vertx.core.impl.ContextInternal.dispatch(ContextInternal.java:279) at io.vertx.core.impl.ContextImpl.lambda$internalExecuteBlocking$2(ContextImpl.java:210) at io.opentelemetry.javaagent.bootstrap.executors.ContextPropagatingRunnable.run(ContextPropagatingRunnable.java:37)

During debugging I found out that grpc context is being forked all the time despite the new independent grpc call. Adding io.opentelemetry.context.Context.current().asContextElement() fixes the issue but having kotlinx-coroutines module inside javaagent should be enough

Steps to reproduce

`class GrpcClientKotlinStub {

companion object {
    val vertx = Vertx.vertx()
    val openTelemetryContext = io.opentelemetry.context.Context.current().asContextElement()

    @JvmStatic
    fun main(): Unit = runBlocking(vertx.blockingDispatcher(false)) {

        val channel = ManagedChannelBuilder
            .forAddress("localhost", 8080)
            .usePlaintext()
            .build()
        val count = AtomicInteger()

        (1..600).map { n ->
            val stub = BofficerApiGrpcKt.BofficerApiCoroutineStub(channel)
            println(n)
            stub.hello(HelloRequest.getDefaultInstance())
                .collect {
                    count.incrementAndGet()
                }
            val context2 = Context.current()
            val generationField2 = Context::class.java.getDeclaredField("generation")
            generationField2.isAccessible = true
            println("generation after: ${generationField2.get(context2) as Int}")
        }
        println("after request, got ${count.get()} items")
        channel.shutdown()
    }
}

}

fun main() {
GrpcClientKotlinStub.main()
}`

Expected behavior

No problem with making grpc call using kotlin coroutines and opentelemetry at the same time

Actual behavior

Exception in io.grpc.Context validateGeneration due to a great number of forks

Javaagent or library instrumentation version

2.4.0

Environment

JDK: 17.0.6
OS: Ubuntu 20.04.3 LTS

Additional context

No response

[KarinaGanieva-sl]

The bug is reproduced only when using Vertx dispatcher

[laurit]

Could you provide a minimal application that reproduces the issue.

[KarinaGanieva-sl]

reproducer https://github.com/KarinaGanieva-sl/opentelemetry-reproducer