OK this wasn't all that scientific but was fun and I tried to have some sort of controls. I picked epoxies I could get locally and had a rating of 3500 PSI or whatever that rating is on the box. I took a 2" piece of steel and sanded it with 100 grit belt. Then I took 3" 3/16" rods and sanded them with 180 grit. I glued them to the metal with 1 1/4" down. I tried to use the same amount of glue as close as I could. The Medium CA was the exception because it wanted to run after it was clamped. That probably proved it's down fall because the dowel popped off when I was inserting the steel bar over that rod.
To do the test I drilled a 3/4" steel rod an inch deep with 3/16" drill. The rod was 5" long. I put this over the dowel and then put a long light weight steel pipe over the 3/4" rod. I had marked the rod every 6" and planned to put weights (in this case a heavy C clamp) on the pipe and move it out to see where they broke. Only 2 survived. All of them except the Devcon plastic welder and the PC-7 failed as soon as I put the pipe on and gently let go. The West system held momentarily while I was reaching for the clamp and then fell. All the others popped off immediately. The PC-7 held up at 6" but failed at 12". The Devcon held up at 24" but failed just as I reached for the clamp to move it further. I had expected the Devcon to fail first since it was called plastic welder.
for my purposes I glued 3/8" dowels into 2 blocks the same distance that I glue my handles to my mirrors. As you can see in the photo all epoxies succeeded. I didn't do a CA test because I know from previous tests it would fail.