rcmgr: debug logs around blocked resources should come from the same place

[MarcoPolo]

I think there's some confusing when users look at the blocked resources log from the resource manager. Specifically the logs don't always line up with the metrics. I think the metrics are more accurate because they are being hooked in via the traces, while the logs are ad-hoc throughout the code.

The fix here is to create a trace reporter that consumes the trace event and logs blocked resource events.