The pace graph for run activities in GC (web) and GCM are both scaled according to your min and max paces during an activity. There is no "true" minimum pace (expressed in minutes:seconds per km) that can correspond to "not moving" (except for infinity seconds per km), so obviously the software has to pick an arbitrary minimum. The problem is that if this minimum is too low (slow), the scale of the graph is zoomed out to the point of uselessness.
GC (web) uses 40:00-50:00/k as the minimum pace, which produces useless pace graphs when you have significant periods of not moving (such as interval workouts with stationary rest). The scale of the graph is zoomed out so far that that 3:30/k and 4:00/k are barely distinguishable.
GCM seems to uses a more reasonable minimum of 10:00/k in similar cases. For comparison, Strava also seems to bottom out around 11:40/k, and I've never had a problem with the graphs on that platform.
It would be great if GC (web) used the same minimum pace as GCM, so graphs would look the same and be useful on both platforms. It would be even better if you could zoom in vertically, effectively choosing your own minimum, to take a closer look at how your pace changed in the range you actually care about. (For example, in an interval workout, I care about when I was going 3:30-4:30/k, not when I was standing still at "40:00/k", and maybe not even when I was jog recovering at 5:30/k)
Example pace graph for interval run on GCM. The graph looks fine and shows me the general trend of my pace (which tells me that I tend to go out wayyyy too fast.)
https://i.postimg.cc/65Fc3t4L/File_001a.png
Same interval run on GC (web). The graph is useless except to see where I was recovering. (You may want to click to zoom and see it at full size).
https://i.postimg.cc/gjQVnHjk/gc-web.png
(I would love to have posted these images inline, but the forum isn't cooperating).