Query suggestion performance time

Elastic Search
1

I have a meta engine that contains 1.8 M docs im assuming that the size of the engine is around 1.5GB.

i tried to tune as much as i can via UI result setting and adding gzip compression.
The best result i got was ~200ms but most of them are above ~750ms.

Im thinking of adding more shards and maybe split the data so smaller engines as some post suggested .

Any other suggestions ?

2

Check out this article on some tips for tuning:

How many shards/replicaes do you currently have for your index? How many data nodes are there? What kind of searches are you doing?

These are all really important questions to consider. Elasticsearch is super easy to use and setup but can be a real pain to optimize correctly :stuck_out_tongue:

3

I created an engine via UI ,
I cant figure out how to configure shards like normal index...

I have 3 node cluster with 2GB RAM on elastic cloud.

the search is pretty simple i have an "Location" engine and im looking for field "city"

{
"query": "san d",
"types": {
"documents": {
"fields": [
"city"
]
}
},
"size": 5
}