Metrics based insight

Automatically digest monitoring data in some way to provide insight into cluster use (i.e, read/write peak times, most frequent performance bottleneck, …).