MapR-DB splits any region that hits 6GB to keep them small. So, as I throw data into MapR-DB (via OpenTSDB), my regions keep splitting.
I have 17 regions after 1 day, and plan to store 60 days' data which would be ~1,000 regions. Is having 1,000 regions on 8 nodes viable from a MapR-DB perspective? Will I still have performance?
Note that I just have 8 MapR nodes that are all pretty hefty.