we hit with data exports after the replica failover

We ran into this around data exports after the replica failover.

Retries arrived out of order and our fixtures only modeled the happy path.

Sleeping in tests hid the races and slowed CI.

Replaying a small set of recorded payloads found the idempotency bug.

We asserted on stored keys, not on elapsed time.

The suite is shorter and finally recreates the incident we cared about.