Is there a restriction for column families per cluster / keys beyond which cassandra's performance will decrease significantly? I know that we had problems with earlier versions, and for some reason, reducing the number of column families solved the problem.
source share