How would you underengineer it? What would be a barebones setup for observability at the scale of one person with a few servers running at most a dozen different scripts?

I would like to make sure that a few recurrent jobs run fine, but by pushing a status instead of polling it.