Search query perfomance issue

Hi All,

I hope you can help me out and point me to the right direction.
We have a three node cluster, and a Bucket with around 100000 documents, we are making simple search for retrieval but from our point of view it takes quite a long time around 1 minute for 100000 documents and 50 MB of data

As we noticed the problem most probably is about data transfer due to the fact that the same request takes 10 times less if we select only IDs, so response size is smaller around 4MB

select META().id from configuration data where type=“person” limit 100000

Could you suggest any way to increase perfomance ?

Best regards
Roman

Data transfer takes time.
It depends on exactly what query is trying to do. If you want get whole document of given type and want to get 100K documents of size 50MB will recommend the following.

CREATE INDEX ix1 ON configuration(type);

FROM the SDKs issue the following query
SELECT META().id FROM configuration WHERE type = “person” LIMIT 100000;
For each document get the document directly from the Data Node asynchronously.
This avoids extra processing by query node avoids 2-hops of 50MB data (i.e Datat Node —> query Service----> Client)