Hello,
I'm investigating an issue raised by one of our user. It seems that after a certain amount of time, a CancelledKeyException is thrown in the NioLoop which stops the NioLoop thread and after this every 10 seconds, this exception appears :
ERROR c.r.c.impl.ForgivingExceptionHandler - Caught an exception during connection recovery!
java.util.concurrent.TimeoutException: null
at com.rabbitmq.utility.BlockingCell.get(BlockingCell.java:77)
The recovery connection loop creates each time a set of object like RecoveryAwareAMQConnection, SocketChannelFrameHandler, SocketChannelFrameHandlerState ... which lead to the memory issue.
We are using the amqp-client 5.14.2 with on top of it the reactor-rabbit in version 1.5.6.