Can you provide Kamanja performance statistics for a use case where 16 billion input messages(in a day) are aggregated over 4 attributes and produce a report. (kind of a group by in a relational database). Let us say, we have a 4 node cluster where each node has 32 cores.
I have seen kamanja performance at 17,000 messages /sec on a 2 node SSD cluster with 2 cores each. However, this was was done with 2 un-optimized models running. I’m positive it could have done better with fine tuning.