The TCP proxy feature can be used for non HTTP traffic. I'd suggest using in an transparent proxy environment, like transparent router or layer 2 mode -- basically in a way such that the original destination is left in tact. That transparent traffic is routed to specific TCP proxy ports on the MWG, then MWG is used as an egress point for that traffic -- no deep inspection is applied, only source, destination based rules could be applied (this includes URL GTI based filtering as well).
If the TCP proxy is not used in a transparent mode, then rules must be created to help route the traffic to the places it needs to go.
Let me know if you had any use cases in mind.
How can I imagine?
If I define a port under the settings, If I define a port under the settings, will only the traffic for this port be passed to the policy engine?
Do I have to build a specific policy in the policy because of the ports?
What happens if I try to intercept this traffic?
Is the traffic then forwarded with the address of the proxy to the server (with the corresponding port)?
Is there a best practice or example setup for TCP proxy somewhere?