Improve back-pressure reporting
Ron has noticed during testing that the system does not report back-pressure in cases where the Fragment generators are producing data at one rate, and DataLoggers are consuming it at a slower rate. As this is likely a more common situation than what the current reporting handles (the downstream stops accepting data entirely for some time interval), we should investigate ways to improve the reporting of back-pressure. Warning messages caused by back-pressure should report the location of any bottlenecks (e.g. "EventBuilder1 has detected back-pressure from downstream!").