With 3.0? Yeah, there was some of that. And still could be. Fact is, if the GPS constellation goes for a crap, and the GPS position estimate goes wandering off... there's not much we can do. It's going to follow. Any system will. However, I know that they did some improvements during the Beta phase of 3.0 that tried to limit that.
I had a few hits myself early on, but haven't had any lately. Actually, that's not true. Last night while I was testing auto landings, 1 out of my ~10 attempts, it was drifting sideways as it touched down. Just slowly. That would have to have been a gps glitch.
I think it won't be too long until we have the Optical Flow sensor working in concert with the INAV and GPS. That will give us some ground-truth that will help when at low altitudes. Should solve the problem almost completely. You'll still have GPS position errors, the best you can expect is ~3m accuracy in absolute position. But you won't get any short term drifting.
But I don't think that will happen on APM2. PX4 only. The APM2 is now running flat out. It's completely redlined, all the processor resources are used up. In fact, I found out a couple weeks ago that the for TradHeli, we don't even have any more program space. I had to make it not-compile-in the OpFlow code to make room.