we have this issue going on where we have been seeing this particular DNode exception, and when it happens we are seeing high CPU utilization, some nodes are reaching up to 300<. We had to restart the application every time. Can someone help us figure the cause. we also turned on logging level to INFO inside SMA for those particular couple of nodes. Below is the exception and some logs,
Cassandra.service.GCIInspector WARN, Heap is full. You may need to reduce or memtable cache sizes.Cassandra will now flush up the two largest memtables. Adjust flush target memtables in yaml file at threshold if want to do this by cassandra automatically.