I stopped doing row-by-row imports once they started hammering the DB. In Migration (unique index), I added a real uniqueness guarantee on provider + uid, because bulk writes only stay correct if the database can enforce constraints. Then, in Bulk ups

I reached for idempotency the moment retries started duplicating side effects. In Advisory lock helper, I generate a deterministic lock ID and use pg_try_advisory_lock to ensure only one worker owns the critical section; the ensure block always calls

I used a transactional outbox when I needed my database write and my event publish to succeed or fail together. In OutboxEvent model, I treated the outbox like a queue: a durable row per event, a dedupe_key for idempotency, and a ready scope that pull