Our cluster nodes have 16 cores and 24GiB. Before expanding the cluster, we had a solid 600% CPU usage (day & night) on all nodes, with a reported 400 IOPs during the day, with 2-3ms IOwait. Only after the cluster expansion to 10 nodes, CPU usage went eventually down during off-hours, still with spikes to 600% during business hours, but not solid 600% like before. Wondering how we can tune-up the badger and his frineds, allocated CPU and memory wise, to finish faster the compaction 