I'm not sure if it's coincidence or you're doing stuff behind the scenes (I imagine the latter) but I have noticed an increase in normal tps on Node 1 since this started last night. Not sure if it's this or not, but felt it was worth mentioning.
I did write some profiling tool for AE2 ME Systems and let it broadcast the most heavy ones on the chat. Within a day the top 4-5 systems got optimized at rates of up to 2000%. The chat bridge runs 99% async. Java was updated which reset my policy adjustment. Should work again with the next server restarts.