It's a bit hard to tell just by looking at the plot what's going on. If you send me the data you're using I could provide a better answer.
That being said, there's a couple of things you can do to try to tweak the Rg if you think the nose isn't real. First, you could try adjusting the averaging window size in the settings, either up or down, to see if that smooths things out (I wouldn't go too high, certainly no more than 20). Second, you do sometimes get noise on the edge of peaks where there's low signal. To remove those points you can adjust the intensity threshold for calculating Rg. Select to the Options->Advanced Options menu item, then in the window that opens go to the Series panel and set the intensity ratio to something higher (1.05 or 1.1, for example), which will use less of the low intensity data around the edges of the peak.
As far as the buffer region goes, you can either use the automated buffer region selection, or you could adjust your manually chosen region to exclude the problematic frames. If you think there isn't really a problem then you can proceed with your selected region. Again, without seeing the data, and how many frames RAW is complaining about, it's a bit hard to say. One or two outliers out of 100 isn't a big deal, but if it thinks there's problems with half the selected buffer frames you should probably investigate further and see what's causing that.
All the best.
Jesse Hopkins, PhD
BioCAT, Sector 18
Advanced Photon Source