So I hear puri.sm was not reachable due to DNS misadministration.
I did not notice.
I run nscd on my laptop and have set persistent host cache and positive-time-to-live value for hosts service to 8 days. (Yes, I’m ignoring problems this creates with dyn DNS hosts, this was never an issue for me). So the whole event passed by while I had the target coordinates in cache.
Maybe Purism would consider extending TTL of their glue records to some large value, chances are this could significantly mitigate DNS outage issues.