There's actually a new (experimental) standard to do this, called L4S. It mandates that the high priority lane apps use a more responsive congestion control alg

There's actually a new (experimental) standard to do this, called L4S. It mandates that the high priority lane apps use a more responsive congestion control algorithm so that they build a much smaller queue (not just on average, avoiding occasional spikes too)