I want to get benefits of ILM (hot, warm, cold) features + the fixed size of indexes without changing the date/time index. Is this possible ?
E.g. I have an index events-2021-03-23 and I want ILM to create a new index like events-2021-03-23-00001, events-2021-03-23-00002 when the shards are all equal at 50GB.
Why I wanna do this?
We're ingesting 1TB+ data per day. Time based index means we query only for the limited time. With long-term retentions, this is a lot of data and I think its best if it stays in time-series format.
If I use ILM and store indexes as firewall-00001,firewall-00002 , will it give result in the same time If I use date range query on firewall-* compared to specifying the indexes firewall-2021-03-23,firewall-2021-03-24 ?
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.