Just as a general observation, there is a third possible solution depending on
the number of columns involved per K1 (which looks reasonable based on your
numbers) but also on how much contention there is for updates to particular
(K1, K2) values (which are always unique partitions in your optio
Hi,
Cassandra schema change is the performance painpoint for us, since it's the
global information across the entire cluster. Our production cassandra
cluster consists of a lot of sets of column families, which totals 1000
column families, and 38301 columns, which sum up to 3.2MB.
We have a data m
