I'd like to modify the snapshot to set lock_wait_timeout equal to the value of "snapshot.lock.timeout.ms", defined in RelationalDatabaseConnectorConfig. This will allow the connector to wait to acquire a lock on the database during snapshot, rather than failing right away.

If there is an existing exclusive lock on the database during the snapshot steps, a connector with default configuration will fail on lock wait timeout (LWTO), while running

FLUSH TABLES WITH READ LOCK;

link
and then later exit on

FLUSH TABLES [table_list] WITH READ LOCK;

link

I've attached logs from the failure described above to the issue.