Memory Leak in Quarkus REST Service if request failed with timeout.

[Urmel2005]

Describe the bug

we use in our services Quarkus to deliver REST API 's .
Unfortunately we detected a memory leak when the request failed for example caused by timeouts.

Used versions:
Quarkus 3.18.4
javaJDK 17.0.5
mvn 3.8.8
quarkus-resteasy-jackson 3.18.4

I attached a small demo Quarkus Service together with a python script called reproducer_client.py.

Demo project to reproduce memory leak.

1. Start the quarkus service with: mvn quarkus:dev
2. Start the python script with: python reproducer_client.py
   The effect only occurs if the client closes the connection before having received a response. (We use a timeout to mimic that)
   If the requests are successful, the memory leak does not occur.
3. Watch Buffer Pools with VisualVM
   Pool count increases when the load is high enough.
   Allocated buffer pools are never released and not always properly re-used.

The memory leak does not occur if the request is successful. It only occurs with aborted requests.
We provide a service with maps and such canceled requests can occur very frequently.

reproducer.zip

Expected behavior

No response

Actual behavior

No response

How to Reproduce?

No response

Output of uname -a or ver

No response

Output of java -version

No response

Mandrel or GraalVM version (if different from Java)

No response

Quarkus version or git rev

No response

Build tool (ie. output of mvnw --version or gradlew --version)

No response

Additional information

No response

[quarkus-bot]

/cc @Karm (native-image), @galderz (native-image), @zakkak (native-image)

[suarkf]

Thank you for the quick fix.
Would it be possible to include this fix also in the current LTS release 3.15.X?

[geoand]

Yes, we will