Context

Existing fast truncate Python tests cover deterministic scenarios with small key sets and fixed timestamps. There is no test that exercises fast truncate correctness under high-volume, randomised workloads across multiple step-up/step-down cycles.

Definition of Done

Add a new Python stress test that:

• Runs many iterations of: follower receives a randomised mix of inserts, point removes, and range truncates at varying timestamps → step-up → assert all keys match expected state → step back down
• Uses a large key space to stress the truncate list and drain logic
• Issues multiple concurrent truncates per round at varying timestamps, interleaved with inserts and removes within truncated ranges
• Tracks expected key state in a Python reference dict and asserts the live DB matches after each step-up