We're building a system using OpenTSDB on top of MapR-DB and it is working quite well.
Before getting too invested though, we wanted to understand our options. We're keen to use MapR-DB for its speed and simplicity. We're also pretty excited about Apache Phoenix (it would make the target system much more powerful and flexible). Will the two work together in the future like MapR-DB and OpenTSDB do?
I found this related question about doing push-downs from Spark to Phoenix - Is Phoenix supported on MaprDb. That would be useful too; but we would be more interested in full support.
We can design the system to be ported from OpenTSDB to Phoenix if we know Phoenix will eventually be available on this platform. We're not particularly interested in using Drill for this purpose and from the outside perspective, it looks like all your SQL eggs are in that basket .
- John Humphreys