I am thinking of having 2 indexer clusters in my environment:
1. “Raw data” cluster, which receives data from windows event forwarders & other “external” connectors.
2. Summary cluster, which receives data from search heads, after those summarized it and took out only part of the “raw data” from cluster 1.
I was wondered whether this is the best solution to my problem, as I want to summarize the data to keep it searchable, which is not possible with the amounts of raw data I have, but still let the users use the “raw data” on real time, so both clusters are needed to be searched.
Is separating the clusters a good idea? Maybe it would be better to use 1 cluster for both purposes, using the same hardware?