maxscale doesn't properly close connections for TCP health check probes

In a cloud deployment in AWS, we use AWS load balancer to send traffic to maxscale.
In absence of traffic, we noticed that the TCP health check probes from AWS LB
were not closed correctly on maxscale side and were stacking up in CLOSE_WAIT state.
These connections were also visible in list session command.

Downgrading to 23.02.8 fixed the issue.