In tracking down some failures on upstream tests #7710 I realized that dask.dataframe should implement numeric_only for reductions like prod, sum and so on. This has some extra urgency now since pandas 1.3.0 will raise a warning if the numeric_only kwarg is not explicitly set.

FutureWarning: Dropping of nuisance columns in DataFrame reductions (with 'numeric_only=None') is deprecated; in a future version this will raise TypeError.  Select only valid columns before calling the reduction.

Ping @kkraus14 and @quasiben to check if cudf needs to implement this kwarg as well.