You are providing support for a Python application running on Compute Engine in production. Recently, there have been reports about the application's slow response times. To investigate how requests are moving through your entire application, what should you do?