@reberhar So Steve, I have a question.
We use Comcast which frequently has significant latency. During the latency times I note that DNS requests backup, causing latency reports because dpinger requests get lost in the pile. Buffer bloat, of course can be a problem. I have read articles about that mentioning that our desire to maintain network requests in a linear fashion might sometimes be misplaced.
So I use priq on my shaper. Is there someway to make that queue smaller? I do note that at the time of the recent problems there is no signficant latency reflected in the gui on the WAN. The logs do reflect very recent latency however. This doesn´t seem like it can be the problem because localhost is resolving DNS requests from the Diagnostics Lookup when I do the testing. And reqular non DNS traffic does not seem to be affected. Stuff that is already in the state table continues to work.
I am just trying to think outside of the box.
