66.249.90.x and 66.249.91.x were indeed blocked.
Ahh... that explains a lot!
I don't know how among many possible ways that those addresses got on the blocking list, as it was too long ago for the current logs to reflect it.
Maybe there was "a lot" of traffic? Possibly also "a lot" in terms of those days.
But of course everyone running a website on an IPIP tunneled ampr.org site has some responsibility in this. Make sure when you have areas with lots of data, those large files are not indexed. This can be done using robots.txt files, headers in the page content, etc.
E.g. you run a site with equipment schematics. You have some text pages with indexes and a lot of huge PDF files with the scanned schematics themselves. It is not difficult to make Google (and other crawlers) index only the text index files and not the PDFs.
Or you have a local amateur group site and it has lots of photographs and maybe even video of the fieldday or other events. It is possible to make the huge 30-megapixel photographs and the video not being indexed and only index the text content and maybe the thumbnails.
When this is done in a responsible manner, indexing the websites that are behind IPIP tunnels should not cause much more "useless traffic" than there already is due to jerks like shodan.io, stretchoid.com and the like. (those are scanning the entire IP range, not just websites that have been announced to Google or are linked from other sites)
Rob