Hi everyone, I have at the moment a SolrCloud cluster with 3 servers and few different collections with all of them replicated to the 3 servers without sharding. I am using streaming expressions retrieving and joining data from these collections so everyhting happens on indivial servers as all data are sitting together in the same server. This expressions are getting pretty complex and we had to scale vertically the instance types on AWS a few times.
I am now trying to split the cluster using sharding and having more servers with smaller instance types to try to see wher we are able to scale horizzontally. First thing I noticed is there is a huge CPU spike when sharing from 1 shard to 8 shards. Is this something which makes sense? I know now we have to move data across servers when on 1 shard everyhitng happened on same server. Why when shardin we have such a big CPU usage increase? Thanks a lot. Sergio Maroto
