Yes, I agree, the only difference is in the Internet for transport between sites. This suggests to me that maybe the problem is MTU related. Is it possible that the Internet connected routers have extra encapsulation like PPPoE or another VPN header?
One thing that I am not sure of is that in the diagram and router configs they establish the VPN over the 192.168.1.x private addresses. I understand this will work in the lab, but if this is the production config as well, then it seems this ipsec tunnel is running over a second tunnel (e.g. Another ipsec tunnel or gre/l2tp). In this scenario the extra encapsulation can definitely contribute to MTU problems. One ipsec encapsulation would be fine though.
The step 1 would be to bypass all traffic. Doing this will pass through all traffic untouched. All traffic should work across the VPN. If not, then there is still some configuration missing for the VPN. Once the two sites can communicate without any issues without Wanos, then we know it is ready to start with the wanop configs.
As a last resort you can try UDP Encapsulation, which is enabled in /tce/etc/wanos/wanos.conf UDPENCAP=Enable
This is required on both sides and Configure > Reset to apply the change. If there are MTU issues, the routers should fragment the packets and Wanos will try to reassemble them. This is not ideal and not recommended as the final solution, but can help in the troubleshooting process.