A service restart solved the issue. The incident started following a short scheduled maintenance of our database provider, after which a service didn't recover.

10% of scripts fail to execute (both syncs & actions).