New Datalab issue type called something like null that checks features for rows that are entirely missing / null values (across all columns).

Those rows should get flagged as is_null_issue.
The quality score for each row can be the fraction of features which are missing in that row.

Make sure this issue check does not waste compute time if it is irrelevant, ie. first check that there even exist any missing values in the features at all before proceeding further.