Get a Protectli box with a very good CPU, 4 Ethernet ports, install OPNSense, and have at it.
I'm surprised the UDM Pro cannot route 2x 1 Gbps on two WANs. I thought it was rated higher than that.
Your test might be running into the 1 Gbps limit backplane problem on the built in switch for UDM Pros.
Maybe he's just testing something temporarily. Like something he bought is having trouble connecting, probably because he's using the same SSID for 2.4 and 5 GHz, but the shitty IoT device can't handle that. So he's messing around, trying to get it to connect.
Unless he's a good friend, I'd just ignore it. You make a comment to the typical person about his WiFi, and he'll become super paranoid about why you even noticed.
Then we'll get a post on this sub from him on how his "creepy neighbor" is hacking his hidden SSID and 60 key password. And now his IoT lightbulbs are now dimming on their own.
And so goes the flow of this sub.
The bitrate of 4K from streaming services like Disney and Netflix is much, much lower than your UHD Blu-ray rips. They recommend having a 16-25 Mbps connection to stream 4K, but the average bitrate is even lower. It's closer to 6-8 Mbits. They just recommend a higher Internet connection because of how streaming works (small bursts of higher rates with a lot of idle time in between).
You can calculate it accurately by just downloading the movie (if the streamer lets you, like premium subscriptions do) to see the file size, and then dividing that size by the length of the movie in seconds. That will give you the average bits per second by definition. You'll be surprised how low it is, because streamers use compression, while "pure" UHD Blu-ray avoids compression to satisfy purists.
As to how much data a streamer uses, it's immense. It's a huge chunk of the data on the Internet at any given time, with estimates in 40-60% range for all the streamers in aggregate. Look into "Content Delivery Networks" (CDNs) to see how it's delivered on a global scale. It's actually very impressive.
There is no set answer, because everyone's environment is different. You'll just need to test it for yourself and see.
First, do speed tests with and without the secondary mesh node. Run a dozen in each configuration to get a usable sample size. Use different speed test sites too.
Then do a continuous ping test to your Default Gateway (your router's LAN address, for example 192.168.1.1 is common, but just check) with and without the secondary mesh node. Run that test for an hour each or more during busy network times, like in the evenings. Compare the results.
Then see which you prefer.
The placement of your mesh nodes, test computer, how busy your WiFi is in general, and layout of your home will determine test results. So there is no set answer.
I will say that, with wireless backhaul, you should just use the least number of nodes you need for full coverage. Four seems excessive. Most homes need only two mesh units to cover everything, three for bigger or unusually shaped houses.
With your setup, if two devices want to communicate, and their ports and the ports on the switches they connect to all supports 10G, then they'll communicate at 10G.
If any of the ports is 1G, even if every other port is 10G, it'll drop down to 1G for that particular communication pathway. That drop down does not "spread" to other pathways.
Having a 1G device plugged into a 10G switch does NOT affect anything else on that switch. Each connection has the "right" to connect at 10G as long as everything along the communication pathway supports it, and is not affected by other concurrent connections that are happening alongside it. Switches can compartmentalize each connection as its own.