or it has some issues?
Yes, but you have to care with data model to avoid big partitions, because reading is restricted by some rules
we dealing with it on cassandra too (we stored all history of changes per partition id - when someone did much updates - it grown and could make serious issues on cluster), but current data model is flat - 1 document per partition
We use big partitions, I think with flat data modeling you will need up to 20 nodes
Обсуждают сегодня