I’ve submitted a PR that sends out a periodic e-mail to users who’s scheduled queries are failing and they have no indication of it. You can read the specifics in the issue, but the gist of it is - these reports are sent at most once an hour, and for each error, they show the amount of times it happened since the last report. If an error fails too much (say, 100 times), the failure is no longer reported. An indication that reports are about to stop is given prior to that.
At first, I thought this should be a text-only e-mail, but following @ranbena’s comment and looking at other failure reports, for example, Sentry’s, I get a feeling we should make our reports better than this piece of :