Thanks Alex for the reply, sorry if i wasn't clear about the rollover.
I am basically rolling over the index based on time/doc count like this:
curl -X POST 'http://localhost:9200/data_collector_index/_rollover?pretty=true' -d '
I originally thought that schema will be copied from the previous schema but later realized that i had to use templates which solved my problem.
I do have one small question, regarding the templates:
should i be defining all the refresh_interval, number_of_shards, number_of_replicas, analyzers and mappings in the template? if so what would be included at the index creation time? just a blank create script?