Active TCP Connection run into "timeout client"

Hi, I found a solution (set timeout tunnel) but still want to understand.
I have a the API of an lxd-server behind HAproxy 2.4, in TCP-Mode.
If push some large data from the Client to this server, the connection breaks.
Some test and I could confirm its always after the “timeout client”.
Setting “timeout tunnel” or increase the client timeout, and the data transfer works well.

The main part of my hyproxy.cfg (and its a playground system, so its not a “good” config :smiley: )

frontend fe_tcp_443
  mode tcp
  option tcplog
  bind :::443 v4v6
  tcp-request inspect-delay 5s
  tcp-request content accept if { req.ssl_hello_type 1 }
  acl sshtls req_ssl_sni -i ssh.07q.de
  acl lxc req_ssl_sni -i lxc.test.07q.de lxd-test
  use_backend lxc if lxc
  acl ssh_payload req.payload(0,7) -m str "SSH-2.0"
  tcp-request content accept if ssh_payload
  acl ssl req.ssl_hello_type 1
  use_backend ssh if ssh_payload
  use_backend terminate_tls if sshtls
  use_backend internal if ssl !sshtls !lxc
  default_backend checkmkagent


backend internal
   mode tcp
   server fe ::1:8080 send-proxy-v2

backend terminate_tls
   mode tcp
   server fe ::1:8081

backend lxc
 mode tcp
 option tcplog
 server lxc 10.15.195.1:8443

backend ssh
  mode tcp
  server lh 10.15.195.1:22

backend checkmkagent
  mode tcp
  server checkmk 10.15.195.1:6556
``