Periodically my clients' blogs get hammered by bots (many GETs per second) that have "erroneous" URLs they are requesting. W3 Total Cache handles bots pretty well for pages that exist, but if a page is 404, then the 404 page isn't cached and the attacking bot just overwhelms us because every 404 requires generation of a full page.
Yes, we can create a streamlined 404, and will do that, but ultimately it would be nicer if W3 Total Cache could cache response code 404 pages as well as response code 200 pages. Would save us a lot of stress on the servers.
Any ideas on whether this is feasible?