I think a lot of folks solved this problem at the application level using background workers or job processing services.
For example, Celery with Python or Sidekiq with Ruby has had recurring job support for a really long time. Celery supported this since 2010. You could run distributed jobs (recurring or not) this way without using cron directly or requiring a custom DB extension.
IMO it feels much more natural to solve this problem at the app level with a background worker since you can run any task you want, not just DB queries. This also has the added benefit of letting you use your DB library of choice (ORM, etc.) since you can run any code you want within the job.
That caveat makes a big difference. If you're making an API call like sending an email, you really don't want it to run twice.
In the payment processing company I'm working at we deal with that independently. We set some data on Redis with a TTL to signal some request has already been made in the last X minutes, and if the same request arrives again during the same period the cached response is served