Hi Joon! Yes, your RAID configuration will help to protect from downtime due to drive failure, but indexer clustering with appropriate replication factors can ensure that indexed data is available elsewhere in the cluster if a whole machine breaks down or is undergoing maintenance.
The default of 3 replicated (raw data) copies and 2 searchable copies guarantees that you can withstand an outage of 2 machines in your cluster without data loss. Searchable copies are special; they can be re-created from replicated copies but that takes time so if uptime is critical you need at least 2 searchable copies in your cluster. Note that the data is replicated on a per-bucket basis.
The defaults are probably reasonable for most environments. Note that increasing the number of searchable copies in particular will cost a lot of space and probably won't deliver much benefit.
Finally, RAID and indexer clustering are not substitutes for backups. Among other things, if you make a mistake in index retention policies or maybe the delete command, your data could be toast without backups.
