Histograms for run/wait times on the various networking reactor threads are included in FTDC, but there isn't a way to observe this data in serverStatus. This information could be useful to see in dashboards and possibly for future mongotune policies, so we should introduce new serverStatus metrics for these.

To cut back on bloat and also to make the metrics more immediately usable, we don't need to include them as histograms. Individual metrics tracking a moving average / total time should be sufficient.