I am currently looking into the rollover feature in elasticsearch.
At the moment i have timebased indexes from filebeats, eg: filebeat.2017-25-05
I would actually like to still keep filebeat create a new daily index and the the rollover to split these when they reach a certain amount of documents.
But cant figure out from the documentation if this is possible?
Something like this? While it is possible, it is not automatic. You would have to use some process to make the daily indices/aliases before the rollover time came, so that Beats could just point to it without having to think about it, otherwise Beats would create a new index named daily-YYYY.MM.DD-1 instead of pointing at a rollover-capable alias.
You can set indices to roll over based on document count and/or time, so you could set it up with a pure rolling index and still make sure no index contains data for more than one day if that is the granularity you need in order to manage retention.
The rollover mechanism controls the underlying index names and Logstash will be indexing into a single write alias. Look at this blog post and the documentation for more details.
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.