As I see this it works when there is behind the scenes backup schedule of database (because once a day isn’t enough, is it?) and then inbuild warningsystem is lagging behind, max 24 hours. But it would work as early warning system if Discourse is down because of database failing, but because of caching users don’t see it right away.
Thanks. If there’s no way to set a “hook” after a backup has been completed, then I think the idea of reverse engineering the discourse API to find the most recent backup, is probably the way to go, and then we have full control over what to do if a backup hasn’t worked… but if a hook (web or shell command) could be added post-backup, that would be ideal.