Search nodes requirement for small amount of data present in database

Recently I come to know that we can introduce search nodes to improve performance. however data present in our cluster is not much but we are using some $regex operations to get some required results. it results into slow query. we can’t use search index(text) because we need to search by using any string or digits or alphabets.
so basically I want to know what is exact need and impact of search nodes ? can we use that irrespective of data size? if not is there any other way to handle $regex slow responses?

Hi @Shubhada_Patil ,

Search nodes allow you to isolate your search workloads from database workloads. Also, you can scale search resources to meet demanding latency/qps requirements. You can use $search without dedicated search nodes. But we strongly recommend using search nodes in any production environment.

What I would recommend is enabling $search without search nodes to see if it will address your use cases. You can always deploy search nodes at a later date without any impact on your existing indexes.


Hi @Seth_Payne,

Thanks for the quick response.

enabling $search means we have to search substring only right as is supports to text index. but sometimes we just need to use some random alphabets to get documents. i think $regex is only solution for that.
for string or substring search i will definitely go with text index.