You'd be OK with putting a ton of work into something, and then having it cached/ripoffed somewhere else?
Caching it for when a site is overloaded is not ripping it off, just like sending copies of works to the British Library for archival purposes is not ripping the original work off.
Also, the library analogy doesn't work in copyright law. Copyright protects the act of copying, not the act of transferring an already-authorized copy of a work to someone else.
Perhaps a simple GET to the URL ever 60s. If you receive 5 non-200 responses in a row, then the HN link points to the archive.is version. Same method in reverse for bringing it back.
How would one request per minute from a single server DoS the site, let alone DDoS it?