I've been trying to figure out how to take advantage of Traffic Shaper for MONTHS! I finally found a configuration that not only works, it works perfectly for what I wanted. I just figured I'd post this in the hopes it helps someone else, like it helped me.
First off, I initially used buffer bloat sites to measure how well the Traffic Shaper was working. However I found that a bad buffer bloat score can be solely caused by a bad workstation config as well. For example, I move my laptop from a C- to an A+ with the following command;
netsh int tcp set global autotuning=disabled
Anyway, I typically use Teams for work, and a few weeks ago one of the kids was downloading a game and it KILLED my throughput.
The article below solved my issue perfectly!
Thank you OPNSense Team for making such a top notch product, available for us plebs.
And thank you Manuel Laggner for the article.
https://www.laggner.info/posts/opnsense-traffic-shaping/
The status screen.
I'm glad your issue is resolved but I'm inclined to believe that it's FQ_CoDel itself which made the difference for you and that the weights aren't really contributing in this setup. It's been said (https://forum.opnsense.org/index.php?topic=36602.msg178873#msg178873) on these forums that weighted queues don't work with FQ_CoDel.
I might be misunderstanding something, though, and happy to be wrong.
As you rely on your internet for business I certainly wouldn't push for you to change anything, but if you have the time/patience you could look over the thread below and the documentation that resulted from it:
https://forum.opnsense.org/index.php?topic=46990.msg242604#msg242604
https://docs.opnsense.org/manual/how-tos/shaper_control_plane.html
Basically in that setup the data plane is using FQ_CoDel (with equally weighted queues set to 100, since they don't apply) and the control plane is using weighted queues (WFQ or QFQ) where weights do apply, but they are initially set to equal ratios. You could however change them.