I'm trying to prevent the insertion of duplicate records from concurrent transactions.
FOR UPDATE does not seem like a good candidate because the records are not updated, so there is nothing to lock. Field-based
unique constraint does not seem like a good candidate because the criteria are a bit more complex.
Unique-constrained formula fields are reasonable candidates, but feel a bit clunky.
Do duplicate rules work at the database-level, similar to unique constraint? Do they prevent insertion of duplicate records across concurrent transactions?