Our 11g RAC cluster has two nodes, each with 8 CPU cores and a memory target of 16GB. System RAM is 64GB (don't ask me why the memory target is just 16GB, I wasn't there when that choice was made). The disk is SSD. The CPU load as shown in the Enterprise Manager never exceeds 25% and is usually around 20%, and the RAM occupancy stays around 20%. The IO operations per second (that worry my the most) vary between 2000 and 8000, with an average around 5500. The load is mostly OLTP style. System load is around 1, maybe 1.5.
My question is, can we consider this cluster to be "heavily" loaded? Again my concern is with IO, obviously CPU and RAM are not stressed.
I'm asking because the application occasionally shows signs of sluggishness. The developers refuse to put logging statements that would show how much time a query takes to run ("too many queries"), so I'm left looking for less direct ways to figure this out.