It would be helpful in the case of a recurrent failure to have more debugging information available, especially as it often is the case that access to the system where the failure(s) occur(s) is either difficult or impossible.