How to continue using same trace from main thread in worker thread with Auto Instrumentation

[suman2007]

I have main thread(service thread) and creating 2 separate threads Thread-1 and Thread-2 and trying to send the rest call from both thread, but there is not any Trace Id found from two thread.
Expectations: Trace Id must be same in Thread-1 and Thread-2 which was generated in Main Thread.

[laurit]

https://opentelemetry.io/docs/languages/java/instrumentation/#context-propagation-between-threads

[suman2007]

These manual changes are required in every portion from where a new thread spawned although I am using Autointrumentation.

[laurit]

Auto instrumentation propagates context when using executors listed in

opentelemetry-java-instrumentation/instrumentation/executors/javaagent/src/main/java/io/opentelemetry/javaagent/instrumentation/executors/ExecutorMatchers.java

Lines 47 to 99 in d852ead

	Arrays.asList(
	"akka.actor.ActorSystemImpl$$anon$1",
	"akka.dispatch.BalancingDispatcher",
	"akka.dispatch.Dispatcher",
	"akka.dispatch.Dispatcher$LazyExecutorServiceDelegate",
	"akka.dispatch.ExecutionContexts$sameThreadExecutionContext$",
	"akka.dispatch.forkjoin.ForkJoinPool",
	"akka.dispatch.ForkJoinExecutorConfigurator$AkkaForkJoinPool",
	"akka.dispatch.MessageDispatcher",
	"akka.dispatch.PinnedDispatcher",
	"com.google.common.util.concurrent.AbstractListeningExecutorService",
	"com.google.common.util.concurrent.MoreExecutors$ListeningDecorator",
	"com.google.common.util.concurrent.MoreExecutors$ScheduledListeningDecorator",
	"io.netty.channel.epoll.EpollEventLoop",
	"io.netty.channel.epoll.EpollEventLoopGroup",
	"io.netty.channel.MultithreadEventLoopGroup",
	"io.netty.channel.nio.NioEventLoop",
	"io.netty.channel.nio.NioEventLoopGroup",
	"io.netty.channel.SingleThreadEventLoop",
	"io.netty.util.concurrent.AbstractEventExecutor",
	"io.netty.util.concurrent.AbstractEventExecutorGroup",
	"io.netty.util.concurrent.AbstractScheduledEventExecutor",
	"io.netty.util.concurrent.DefaultEventExecutor",
	"io.netty.util.concurrent.DefaultEventExecutorGroup",
	"io.netty.util.concurrent.GlobalEventExecutor",
	"io.netty.util.concurrent.MultithreadEventExecutorGroup",
	"io.netty.util.concurrent.SingleThreadEventExecutor",
	"java.util.concurrent.AbstractExecutorService",
	"java.util.concurrent.CompletableFuture$ThreadPerTaskExecutor",
	"java.util.concurrent.Executors$DelegatedExecutorService",
	"java.util.concurrent.Executors$FinalizableDelegatedExecutorService",
	"java.util.concurrent.ForkJoinPool",
	"java.util.concurrent.ScheduledThreadPoolExecutor",
	"java.util.concurrent.ThreadPoolExecutor",
	"java.util.concurrent.ThreadPerTaskExecutor",
	"org.apache.tomcat.util.threads.ThreadPoolExecutor",
	"org.eclipse.jetty.util.thread.QueuedThreadPool", // dispatch() covered in the jetty
	// module
	"org.eclipse.jetty.util.thread.ReservedThreadExecutor",
	"org.glassfish.grizzly.threadpool.GrizzlyExecutorService",
	"org.jboss.threads.EnhancedQueueExecutor",
	"org.apache.pekko.dispatch.BalancingDispatcher",
	"org.apache.pekko.dispatch.Dispatcher",
	"org.apache.pekko.dispatch.Dispatcher$LazyExecutorServiceDelegate",
	"org.apache.pekko.dispatch.ExecutionContexts$sameThreadExecutionContext$",
	"org.apache.pekko.dispatch.ForkJoinExecutorConfigurator$PekkoForkJoinPool",
	"org.apache.pekko.dispatch.MessageDispatcher",
	"org.apache.pekko.dispatch.PinnedDispatcher",
	"play.api.libs.streams.Execution$trampoline$",
	"play.shaded.ahc.io.netty.util.concurrent.ThreadPerTaskExecutor",
	"scala.concurrent.forkjoin.ForkJoinPool",
	"scala.concurrent.Future$InternalCallbackExecutor$",
	"scala.concurrent.impl.ExecutionContextImpl"));

When something else is used manual instrumentation may be needed.

[suman2007]

Thanks for the update!!

[suman2007]

Is there any reason that Context Propagation in Auto instrumentation is not supported for the traditional Thread approach or there is plan to incorporate this missing thing in the upcoming release?

[trask]

Check out #2527