Just looking at Screenshot_20230817-170213_Gallery.jpg the side view if the set-up, it could be the camera angle but it doesn’t look like there’s much fall on the first section of 110mm from the gully to the 15° bend. I also notice that there’s a short section of 110mm at right angles to the main pipe being used to prop up the assembly before the rubber connector. Is this how you found it after the installation?Here are some photos. I've tried putting an endoscopic camera down to look for any issues but the bend where it drops down underground presents a challenge I can't get past.
It is worth stressing that a lot of water comes down there from the flat roof and two sides of the main roof. I wonder if it is some strange pressure issue? As I say, works fine as soon as I remove the insert and continues to work fine once I put it back in - with continued flow from downpipe.
Just wondering if you’ve tried sitting a spirit level on this section? As an aside, It always amazes me how little fall is required to do the job with foul waste pipes, but I’ve found with rainwater the more the merrier within reason, as the flow can be very peaky, so what about having as much fall as you can get away with? As a flexible rubber coupler i s present, is it worth approaching the installer to see if the 110mm pipe prop could be replaced with something smaller (lower) to increase the fall? In fact, could you get the installer back in to ask why this issue is occurring?
What’s bothers me is that I’d go for what Ivor said, i.e. lift the cover of the inspection chamber (wherever that is) to see what’s happening there, however, your observations upstream kind of point to the insert itself being an issue. I do wonder if hypothetically, you swapped in a gully of another design, this issue would just go away?