When the scraping fails due to a host error, we should be able to flag this in an appropriate way for the server and the affected repositories.