How to properly configure a cluster of indexers to work ES?

[sh_bolatbekov]

Hello!
We need to implement architecture ES Splunk to 400 GB in clustering (SH, IDX). How we should to count numbers of idx for this capacity? In Splunk docs recommend to use per indexer up to 100Gb. How it will be in IDX clustering? We need data replication on index cluster. How index cluster will replicate 400GB data if Splunk docs recommend to use ES per indexer up to 100Gb?

[richgalloway · Accepted solution]

You will need at least 4 indexers. Note the "up to 100GB" in the recommendation. That means you may get less throughput per indexer depending on search activity and other factors. I suggest 5 indexers.
Replication mainly affects storage. You'll need more disk space to hold the replicated indexes. See https://splunk-sizing.appspot.com/

[LukeB]

Hello,

can you send me a link for the "up to 100GB" recommendation?

[PickleRick]

Firstly - please don't dig up threads from four years ago.

Secondly - there's no document specifying "you can ingest 100GB and you'll be fine" since noone will ever give you 100% guarantee. It's a rough estimate assuming some normal conditions using reference indexers. But you might hit your capacity limits way earlier if your search is heavy.