Thanks for the fast reply. So what you're saying basically is monitoring and tracing when performance is ok, then comparing with the trace when it's not performing as it should?