Stomp acceptor causes address memory rise

[Herbert . Helmstreit]

Hello Community

in an artemis 2.33 cluster with relatively high rate of core messages we 
enabled a stomp acceptor on one broker.

Compared to the other brokers in the cluster that have <1 MiB used Address 
Memory the modified guy goes up to 250 MiB.
There is no stomp client attached. The value of the address memory gauge 
seems to be sensitive to the message rate.
We are using non-persistent messaging with a TTL of 3 min. Byterate*TTL 
could be close to the value of 250 MiB.

Is the Messagebuffer to high? 
{
  &quot;scheme&quot;: &quot;tcp&quot;,
  &quot;tcpReceiveBufferSize&quot;: &quot;1048576&quot;,
  &quot;port&quot;: &quot;61613&quot;,
  &quot;host&quot;: &quot;0.0.0.0&quot;,
  &quot;protocols&quot;: &quot;STOMP&quot;,
  &quot;useEpoll&quot;: &quot;true&quot;,
  &quot;tcpSendBufferSize&quot;: &quot;1048576&quot;
}

Can somebody recommend a measure?

Best Regards

Herbert