Logging, monitoring, profiling, and debugging distributed systems, as discussed in Chapter 7, Testing and Debugging Distributed Applications, even today is not an easy task, especially when using languages other than C, C++, and Fortran. There is not much more to say here other than the fact that there is an important vacuum to be filled.
Most medium-to-large teams end up developing their own custom solutions based on log aggregators such as Sentry (https://getsentry.com) and monitoring solutions such as Ganglia (http://ganglia.sourceforge.net).
What would be nice to have are the equivalent of I/O monitoring tools such as Darshan (http://www.mcs.anl.gov/research/projects/darshan/) and distributed profilers such as Allinea MAP (http://www.allinea.com/products/map) for Python distributed applications.