Due to their randomness UUIDs might be sub-optimal for database operations, see https://www.cybertec-postgresql.com/en/unexpected-downsides-of-uuid-keys-in-postgresql/
- we are actually hit by the problem
- we cannot easily change the UUIDs later in production
https://www.ietf.org/archive/id/draft-peabody-dispatch-new-uuid-format-04.html
Not yet decided