I've noticed a few times with libre if I scan and get, say, a 5ish, then shortly after get an 8 then a short while later 5ish. The errant one doesn't show on the graph because it is constructed from averages over 15 minutes. I wonder sometimes whether it's maybe just that glucose isn't that evenly distributed in either blood or interstitial fluid, so maybe the outlier just happened to be a more sugary drop of blood or fluid than it's neighbours. Have no idea whether that's the case, just a thought.