In addition to what @nik9000 said... each index consists of N shards and M replicas. Each shard is equivalent to one Lucene index and each Lucene index can hold roughly 2B documents. From here, you can do your own calculation in terms of how many shards one index should have when you want to have 1) one big index 2) one index per month and so on.
I want to calculate Retention rate of Apache access log data for one month. Can you help me how to get it done. I have been told to use curator , but I don't know how to do .
Can you suggest any better approach or any document / link related to curator, which can be helpful in my case ?