Currently queue of requests to make replication process surviving Nexus restart is being fully rewrite every change. This may result in performance issue. Replace this solution with for example, SQLite DB.