For test purposes we currently have an index containing about 50M docs, 
distributed on a 4 nodes cluster, with 16 shards.
Do you think that drastically increasing the number of shards would help ? 

On Friday, January 31, 2014 10:14:08 AM UTC-8, Binh Ly wrote:
>
> Maxime, forgot to mention, you can also distribute the load out by 
> increasing the shard count and adding more nodes. But precomputing the 
> field is probably the quickest way to improve that performance. Keep in 
> mind that unlike SQL, ES aggregations may return approximate metrics if you 
> have more than 1 shard.
>

-- 
You received this message because you are subscribed to the Google Groups 
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/elasticsearch/21874399-99c9-4c6b-8c76-f856ff95216f%40googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

Reply via email to