Ignoring fields from indexing and updating same fields

Hi everyone,

We are about to release Elasticsearch in production for our time-series analytics data.
We have few concerns about optimizing performance.

By default ES indexes all fields and we want to prevent that. As we understand we can do that with settings "index": false, to those fields by which we don't want to query. Is there more efficient way to do this, like nesting all that data into "enabled": false type of data?

On some data types we do have frequent updates, but those updates are only on fields that doesn't need to be indexed, will ES still reindex all the data each time update happens on these fields?

Is there any way to update field without reindexing?

Ivan Petrovic

You are right. At the moment, updating means a full reindex operation.

You can use index templates and dynamic templates to make sure none of any new fields get indexed by default.

Thanks for sharing dynamic templates, but it's not the solution we are looking for.

It is great that you can make sure that none of new fields get indexed, but we are looking for something a bit different.

Are there any possibilities and plans to improve updates that won't reindex for fields that are not part of the index?

Documents are always indexed as a whole for now.

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.