Can a different service solve my elastic search use cases for cheaper?

Viewed 27

I use elastic search to:

  • Search for items by name
  • Search for items that are hats
  • Search for items that aren't hats
  • Combos of the above

My data looks like this:

{
        "id": string,
        "name": string,
        "owner_user_id": number,
        "moderation_status": string,
        "component_keys": ["Hat", "Clothing", "Render"] # There are hundreds of unique values here
}

My cluster has 104 shards and only uses 93 MB of memory.

I'm a hobby developer, and Elastic is my highest operating expense right now.

1 Answers

Elasticsearch cost depends on the multiple factors, like memory, disk, instance type and number of nodes in the cluster. And Elasticsearch cluster can have minimum 1 to no upper limit specified by Elasticsearch.

In your case I can see you have very less memory footprint so you can simply create a single node Elasticsearch cluster(if you are not using it for production use-case), and just assign the heap memory to 1 GB, this should solve your cost issue.

Apart from that curious to know why you have 104 shards, better to delete the unused shard to claim even more memory.

Hope this helps.

Related