Introduce manual tests with failpoints to stress-test various failure scenarios that may occur during upgrade and downgrade operations while transitioning to database authoritative shards.

While concurrent FSM tests and specialized suites (e.g., stepdowns suites) already cover many failure scenarios, having dedicated manual tests will ensure these cases are consistently run in the CI pipeline.

These tests should specifically target database DDL operations running concurrently with the upgrade/downgrade process, validating system behavior under failure conditions, using specific failpoints.