Node Classification in cluster
Our application is currently deployed with below cluster configuration
11 background nodes , 2 stream nodes and 3 web users.
Now we forsee almost 100% additional load due to new implementation, so we are planning to add additional nodes. As most of the processing in application is background jobs we are planning to add 9 new background nodes. so total background nodes becomes 20.
So now the query is , as we increase the number of background nodes, do you recommend having additional stream node as well ?
In that case may we know how should we plan the total number of stream nodes in such configurations, or just having 2 stream nodes is enough.
Note : Most of the background process we have is via Queue process.