In our earlier attempts to work with JG OLAP, we hit the bottleneck / bounded by scylla virtual token ring count (it appears thats the unit of parallelism while bulk export)
Can you please share your experience while running the bulk export / OLAP on JG directly?
