Nullpointer exception for pekko-instrumentation in version 1.1.0-M1 onward #1352

TjarkoG · 2024-07-25T10:09:59Z

when testing pekko with the version 1.1.0-M1 which will be released soon we've ran into a nullpointer error when running with Kamon + kanela-instrumentation.

to Reproduce:
build.sbt

name := "pekko-quickstart-scala"

version := "1.0"

scalaVersion := "2.13.14"

lazy val pekkoVersion = "1.1.0-M1"

run/fork := true

libraryDependencies ++= Seq(
  "org.apache.pekko" %% "pekko-actor-typed" % pekkoVersion,
  "org.apache.pekko" %% "pekko-http" % pekkoVersion,
  "org.apache.pekko" %% "pekko-stream" % pekkoVersion,
  "ch.qos.logback" % "logback-classic" % "1.5.6",
  "io.kamon" %% "kamon-pekko" % "2.7.3",
  "io.kamon" %% "kamon-system-metrics" % "2.7.3",
  "io.kamon" %% "kamon-prometheus" % "2.7.3"
)

Main.scala:

//#full-example
package com.example

import com.typesafe.config.ConfigFactory
import org.apache.pekko.actor.ActorSystem
import org.apache.pekko.http.scaladsl.Http
import org.apache.pekko.http.scaladsl.client.RequestBuilding.Get
import org.apache.pekko.http.scaladsl.server.Directives

object Main extends Directives {
  def main(args: Array[String]): Unit = {
    val config = ConfigFactory.load()
    implicit val untyped: ActorSystem = ActorSystem.create("Test", config)
    Http()(untyped).singleRequest(Get("https://www.google.com")).onComplete(rsp => println(rsp.map(_.status)))(untyped.dispatcher)
  }
}

when starting with -javaagent:javaagent/kanela-agent.jar (version 1.0.18)

i'm receiving the exception:

org.apache.pekko.actor.ActorInitializationException: pekko://Test/system/IO-TCP/selectors/$a: exception during creation, root cause message: [Cannot invoke "org.apache.pekko.dispatch.DispatcherPrerequisites.settings()" because the return value of "kamon.instrumentation.pekko.instrumentations.DispatcherInfo$HasDispatcherPrerequisites.dispatcherPrerequisites()" is null]
	at org.apache.pekko.actor.ActorCell.create(ActorCell.scala:679)
	at org.apache.pekko.actor.ActorCell.invokeAll$1(ActorCell.scala:523)
	at org.apache.pekko.actor.ActorCell.systemInvoke(ActorCell.scala:545)
	at org.apache.pekko.dispatch.Mailbox.processAllSystemMessages(Mailbox.scala:297)
	at org.apache.pekko.dispatch.Mailbox.run(Mailbox.scala:232)
	at kamon.instrumentation.executor.ExecutorInstrumentation$InstrumentedForkJoinPool$TimingRunnable.run(ExecutorInstrumentation.scala:800)
	at org.apache.pekko.dispatch.ForkJoinExecutorConfigurator$PekkoForkJoinTask.exec(ForkJoinExecutorConfigurator.scala:61)
	at java.base/java.util.concurrent.ForkJoinTask.doExec(ForkJoinTask.java:373)
	at java.base/java.util.concurrent.ForkJoinPool$WorkQueue.topLevelExec(ForkJoinPool.java:1182)
	at java.base/java.util.concurrent.ForkJoinPool.scan(ForkJoinPool.java:1655)
	at java.base/java.util.concurrent.ForkJoinPool.runWorker(ForkJoinPool.java:1622)
	at java.base/java.util.concurrent.ForkJoinWorkerThread.run(ForkJoinWorkerThread.java:165)
Caused by: java.lang.reflect.InvocationTargetException: null
	at java.base/jdk.internal.reflect.NativeConstructorAccessorImpl.newInstance0(Native Method)
	at java.base/jdk.internal.reflect.NativeConstructorAccessorImpl.newInstance(NativeConstructorAccessorImpl.java:77)
	at java.base/jdk.internal.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAccessorImpl.java:45)
	at java.base/java.lang.reflect.Constructor.newInstanceWithCaller(Constructor.java:499)
	at java.base/java.lang.reflect.Constructor.newInstance(Constructor.java:480)
	at org.apache.pekko.util.Reflect$.instantiate(Reflect.scala:82)
	at org.apache.pekko.actor.ArgsReflectConstructor.produce(IndirectActorProducer.scala:111)
	at org.apache.pekko.actor.Props.newActor(Props.scala:236)
	at org.apache.pekko.actor.ActorCell.newActor(ActorCell.scala:626)
	at org.apache.pekko.actor.ActorCell.create(ActorCell.scala:653)
	... 11 common frames omitted
Caused by: java.lang.NullPointerException: Cannot invoke "org.apache.pekko.dispatch.DispatcherPrerequisites.settings()" because the return value of "kamon.instrumentation.pekko.instrumentations.DispatcherInfo$HasDispatcherPrerequisites.dispatcherPrerequisites()" is null
	at kamon.instrumentation.pekko.instrumentations.InstrumentNewExecutorServiceOnPekko$.around(DispatcherInstrumentation.scala:130)
	at kamon.instrumentation.pekko.instrumentations.InstrumentNewExecutorServiceOnPekko.around(DispatcherInstrumentation.scala)
	at org.apache.pekko.dispatch.ThreadPoolConfig$ThreadPoolExecutorServiceFactory.createExecutorService(ThreadPoolBuilder.scala)
	at org.apache.pekko.dispatch.Dispatcher$LazyExecutorServiceDelegate.executor$lzycompute(Dispatcher.scala:54)
	at org.apache.pekko.dispatch.Dispatcher$LazyExecutorServiceDelegate.executor(Dispatcher.scala:54)
	at org.apache.pekko.dispatch.ExecutorServiceDelegate.execute(ThreadPoolBuilder.scala:232)
	at org.apache.pekko.dispatch.ExecutorServiceDelegate.execute$(ThreadPoolBuilder.scala:232)
	at org.apache.pekko.dispatch.Dispatcher$LazyExecutorServiceDelegate.execute(Dispatcher.scala:53)
	at org.apache.pekko.dispatch.Dispatcher.executeTask(Dispatcher.scala:91)
	at org.apache.pekko.dispatch.MessageDispatcher.unbatchedExecute(AbstractDispatcher.scala:177)
	at org.apache.pekko.dispatch.BatchingExecutor.execute(BatchingExecutor.scala:143)
	at org.apache.pekko.dispatch.BatchingExecutor.execute$(BatchingExecutor.scala:134)
	at org.apache.pekko.dispatch.MessageDispatcher.execute(AbstractDispatcher.scala:107)
	at org.apache.pekko.util.SerializedSuspendableExecutionContext.attach(SerializedSuspendableExecutionContext.scala:92)
	at org.apache.pekko.util.SerializedSuspendableExecutionContext.execute(SerializedSuspendableExecutionContext.scala:95)
	at org.apache.pekko.io.SelectionHandler$ChannelRegistryImpl.<init>(SelectionHandler.scala:200)
	at org.apache.pekko.io.SelectionHandler.<init>(SelectionHandler.scala:324)
	... 21 common frames omitted

i've tried to analyze the error and got to the point that
kamon/instrumentation/pekko/instrumentations/DispatcherInstrumentation.scala:164
does not seem to get triggered with pekko 1.1.0-M1 anymore resulting in dispatcherPrerequisites being null in
kamon/instrumentation/pekko/instrumentations/DispatcherInstrumentation.scala:125
but i cannot understand what change in pekko 1.0.3 <-> 1.1.0-M1 made that difference and how to fix it

The text was updated successfully, but these errors were encountered:

pjfanning · 2024-07-26T11:45:44Z

Pekko 1.1.0-M1 is released. It is a milestone release but the jars are all up on Maven Central.

pjfanning · 2024-07-26T14:59:23Z

#1354 PR experiment does not seem to pick up this issue. Could be a gap in test coverage. @TjarkoG could you see if you could add a unit test that reproduces this?

pjfanning · 2024-07-26T20:35:49Z

There are some changes in the Pekko dispatch code in v1.1.0-M1. I wouldn't have thought them major.

https://github.com/apache/pekko/commits/0930982b9a5625e622c4c4c2a6f71dc9917cb336/actor/src/main/scala/org/apache/pekko/dispatch

apache/pekko#485 is the biggest change

TjarkoG · 2024-07-29T13:20:23Z

@pjfanning sry for the late response.
I've looked into it but the whole instrumentation logic is sadly out of my knowledge zone.
Took me way longer to gather the information i put above than im proud to admit :(

pjfanning · 2024-07-29T13:30:51Z

The strange thing is the #1354 CI tests pass for pekko. Any interest in just building your own kamon jars using the branch underlying #1354 ?

TjarkoG · 2024-07-29T13:36:19Z

i can try if this would help with our problem but im not to optimistic that it will help.
the problems with the "normal" jars are also only occurring when starting with the kanela-agent for instrumentation and when using the dispatcher "pekko.io.pinned-dispatcher" from the reference.conf "pekko.io.tcp"
edit:
i've build the jar locally based on #1354 and tried it with the same result

TjarkoG · 2024-08-09T11:00:48Z

@ivantopo do you have any idea how to create tests for that?

pjfanning · 2024-08-10T16:50:46Z

I created https://github.com/pjfanning/kamon-pekko-http-test based on the description here and this runs fine with no issues.

TjarkoG · 2024-08-10T16:54:24Z

Did you start it with the kanela java-agent?
Without it does work but we dont get the instrumentation

pjfanning · 2024-08-10T17:06:35Z

Did you start it with the kanela java-agent? Without it does work but we don't get the instrumentation

Thanks. I hadn't added kanela-agent. The issue happens with it enabled. I updated the sample project. I haven't used Kamon since its early days. I have my own fork that I use myself and it works with Pekko 1.1 (https://github.com/pjfanning/micrometer-pekko).

pjfanning · 2024-08-10T18:51:00Z

I may have messed up the test run but when I tried Scala 3.3.3 and 2.12.19, it seemed ok. Maybe this only happens with Scala 2.13.

Philippus · 2024-09-05T13:23:41Z

I'm getting this issue now with Pekko 1.1.0.

Philippus · 2024-09-05T19:09:21Z

I think it's related to inlining. This commit still works: apache/pekko@d829637. This one doesn't: apache/pekko@0f1db53. And if I disable inlining with the pekko.no.inline setting and compile pekko it works.
I'm using Scala 2.13.14.

pjfanning · 2024-09-05T19:56:10Z

I think it's related to inlining. This commit still works: apache/pekko@d829637. This one doesn't: apache/pekko@0f1db53. And if I disable inlining with the pekko.no.inline setting and compile pekko it works. I'm using Scala 2.13.14.

Seems plausible. Do you see the same behaviour as me - in that Scala 3 and Scala 2.12 work fine but that the NPE pops up with Scala 2.13? The Pekko inlining does not affect the Scala 3 release of Pekko.

Philippus · 2024-09-05T20:04:50Z

I think it's related to inlining. This commit still works: apache/pekko@d829637. This one doesn't: apache/pekko@0f1db53. And if I disable inlining with the pekko.no.inline setting and compile pekko it works. I'm using Scala 2.13.14.

Seems plausible. Do you see the same behaviour as me - in that Scala 3 and Scala 2.12 work fine but that the NPE pops up with Scala 2.13? The Pekko inlining does not affect the Scala 3 release of Pekko.

2.12.19 and 3.5.0 didn't give a NPE.

Philippus · 2024-09-09T12:53:31Z

Is this a bug in Kamon, or a bug in Pekko ?

pjfanning · 2024-09-09T13:02:46Z

Pekko 1.1 uses Scala 2 compiler inlining. Kamon doesn't seem to be able to find the inlined code in Pekko.

leszekgruchala · 2024-09-09T14:21:30Z

I do have the same issue with 2.13.14

pjfanning · 2024-09-09T17:16:17Z

To be clear, the Apache Pekko 1.1 releases won't be modified to suit Kamon. We reserve the right in minor releases to make some changes.
The issue here is that Kanela works a bit like AOP and expects to be able the manipulate Pekko byte code to gather metrics. It looks like the instrumentation logic in Kamon - specify for Pekko dispatchers - needs to be modified to work with Pekko 1.1.

@mdedetrich just adding to you to the issue just in case you have any thoughts on this.

If anyone has any thoughts on hooks that could be added to Pekko 1.2 or 2.0 that makes it easier for Kamon to instrument Pekko then feel free to raise issues or PRs in the Pekko repo(s).

pjfanning · 2024-09-10T00:40:18Z

I created #1361 as a workaround.

pjfanning · 2024-09-10T00:46:53Z

Users hitting this issue should also be able to disable dispatcher instrumentation. See https://github.com/kamon-io/Kamon/blob/124e223de4d7d806b47c76323f5d91edf626fa81/instrumentation/kamon-pekko/src/main/resources/reference.conf and check about excluding dispatchers.

hughsimpson · 2024-09-17T12:06:36Z

Just catching up on all this, been neglecting it over the summer (sorry about that!). The instrumentation is definitely failing because of the inlining (added some comments about this on a draft pr before I realised you guys had already spotted it 😅). If we can get some instrumentation hooks into pekko I'm sure that would mitigate the issue... I will have a poke about with what might work on the pekko side and see if we can find a place of agreement. Almost all the instrumentation should work for scala 2.12, all of it should work for scala 3, so it's really only 2.13 that suffers much as it currently stands; one fix is, thus, to upgrade to scala 3 (although that's clearly not always possible). From my experiments so far, I don't think it's possible to fix this purely on the Kamon side.

mdedetrich · 2024-09-17T12:23:06Z

I defintiely think that instrumentation hooks is the way to go. While inlining may have caused the issue, as @pjfanning pointed out Pekko reserves the right to change the structure of stack/methods as it pleases as long as the methods are not public and so the current implementation of kamon does appear to be brittle for this reason.

If you have some ideas about instrumentation, PR's/design documents are welcome! Supporting this officially via an API is the way to go.

pjfanning · 2024-09-17T12:52:12Z

I debugged a broken test in https://github.com/pjfanning/micrometer-pekko (uses AOP and has similar issues to Kamon) and in this once case that calls to ActorCell.stop were being inlined on one particular call path (Scala 2.13 only). The call path related to stopping a RobinRobinRouter instance that had 5 actor refs to round robin over. What I couldn't work out is why the ActorCell.stop got inlined because it is not marked with @inline annotation. I decompiled some classes and output some stack traces but these made it look like no inlining occurred but the instrumentation is definitely missing the ActorCell.stop call on this call path. The test case works fine with Scala 2.12 and with Pekko 1.0 (Scala 2.12 and Scala 2.13).

pjfanning · 2024-09-17T12:52:57Z

@mdedetrich is there any way to mark a function so it can't be inlined?

Edit - apparently there is a @noinline annotation. https://www.baeldung.com/scala/inline-noinline-annotations

pjfanning · 2024-09-17T12:58:23Z

I raised apache/pekko#1484 - no guarantees that the community will agree to make changes but it is open for discussion.

hughsimpson · 2024-09-17T13:00:16Z

Thanks @pjfanning ! I'll see if I can get a pekko build that passes the suite with a sprinkle of those annotations 🙏

TjarkoG mentioned this issue Jul 26, 2024

in TlsActor flush data to user only after handshake has finished (#1128) apache/pekko#1150

Closed

pjfanning mentioned this issue Jul 26, 2024

Latest pekko milestone #1354

Closed

pjfanning mentioned this issue Sep 10, 2024

make kamon-pekko tolerant of null dispatcherPrerequisites #1361

Merged

pjfanning mentioned this issue Sep 17, 2024

consider marking some functions as @noinline to fix issues with using Kamon apache/pekko#1484

Open

This was referenced Sep 19, 2024

ThreadPoolConfigCopyAdvice often sets null prereqs #1365

Closed

kamon-pekko: find new way to instrument pinned dispatchers #1366

Open

hughsimpson mentioned this issue Sep 20, 2024

Fix downstream kamon instrumentation apache/pekko#1489

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Nullpointer exception for pekko-instrumentation in version 1.1.0-M1 onward #1352

Nullpointer exception for pekko-instrumentation in version 1.1.0-M1 onward #1352

TjarkoG commented Jul 25, 2024

pjfanning commented Jul 26, 2024

pjfanning commented Jul 26, 2024

pjfanning commented Jul 26, 2024

TjarkoG commented Jul 29, 2024 •

edited

Loading

pjfanning commented Jul 29, 2024

TjarkoG commented Jul 29, 2024 •

edited

Loading

TjarkoG commented Aug 9, 2024

pjfanning commented Aug 10, 2024

TjarkoG commented Aug 10, 2024

pjfanning commented Aug 10, 2024

pjfanning commented Aug 10, 2024

Philippus commented Sep 5, 2024 •

edited

Loading

Philippus commented Sep 5, 2024 •

edited

Loading

pjfanning commented Sep 5, 2024 •

edited

Loading

Philippus commented Sep 5, 2024

Philippus commented Sep 9, 2024

pjfanning commented Sep 9, 2024

leszekgruchala commented Sep 9, 2024

pjfanning commented Sep 9, 2024 •

edited

Loading

pjfanning commented Sep 10, 2024

pjfanning commented Sep 10, 2024

hughsimpson commented Sep 17, 2024

mdedetrich commented Sep 17, 2024

pjfanning commented Sep 17, 2024

pjfanning commented Sep 17, 2024 •

edited

Loading

pjfanning commented Sep 17, 2024

hughsimpson commented Sep 17, 2024

Nullpointer exception for pekko-instrumentation in version 1.1.0-M1 onward #1352

Nullpointer exception for pekko-instrumentation in version 1.1.0-M1 onward #1352

Comments

TjarkoG commented Jul 25, 2024

pjfanning commented Jul 26, 2024

pjfanning commented Jul 26, 2024

pjfanning commented Jul 26, 2024

TjarkoG commented Jul 29, 2024 • edited Loading

pjfanning commented Jul 29, 2024

TjarkoG commented Jul 29, 2024 • edited Loading

TjarkoG commented Aug 9, 2024

pjfanning commented Aug 10, 2024

TjarkoG commented Aug 10, 2024

pjfanning commented Aug 10, 2024

pjfanning commented Aug 10, 2024

Philippus commented Sep 5, 2024 • edited Loading

Philippus commented Sep 5, 2024 • edited Loading

pjfanning commented Sep 5, 2024 • edited Loading

Philippus commented Sep 5, 2024

Philippus commented Sep 9, 2024

pjfanning commented Sep 9, 2024

leszekgruchala commented Sep 9, 2024

pjfanning commented Sep 9, 2024 • edited Loading

pjfanning commented Sep 10, 2024

pjfanning commented Sep 10, 2024

hughsimpson commented Sep 17, 2024

mdedetrich commented Sep 17, 2024

pjfanning commented Sep 17, 2024

pjfanning commented Sep 17, 2024 • edited Loading

pjfanning commented Sep 17, 2024

hughsimpson commented Sep 17, 2024

TjarkoG commented Jul 29, 2024 •

edited

Loading

TjarkoG commented Jul 29, 2024 •

edited

Loading

Philippus commented Sep 5, 2024 •

edited

Loading

Philippus commented Sep 5, 2024 •

edited

Loading

pjfanning commented Sep 5, 2024 •

edited

Loading

pjfanning commented Sep 9, 2024 •

edited

Loading

pjfanning commented Sep 17, 2024 •

edited

Loading