Would be nice to see some metrics for, as well as validate our assumption that using compressed buckets reduces WT cache usage.

This is just to get some initial benchmark numbers for the intern presentation. We'll save running the full 2-day long high-cardinality workload until after the intern presentation.