Open search 2.16 performance issues

OpenSearch
1

Versions (relevant - OpenSearch/Dashboard/Server OS/Browser): 2.16

Describe the issue: We migrated the data from opensearch 1.22 to 2.16 ( used snapshot and backfilled the delta update ) . But after everything is done we are seeing the high resp time on the new opensearch cluster (>3X times ) than older open search cluster 1.x , we are using graviton instances (8 vcpus ) and 32 shards config with 32 data nodes . Each shard size is about 5 GB . Any recommendation would be helpful to improve the performance

Configuration: 32 shards and 32 data nodes .

Relevant Logs or Screenshots:

2

Hi @imdbtoimdb,

This article has some “best practices” that might help you optimize your cluster: Sizing Amazon OpenSearch Service domains - Amazon OpenSearch Service. It’s about AWS service but should mostly apply to your local deployment.

Let me know if you have further questions or want to dive deeper into any of the topics.

best,
mj

3

Hi @Mantas ,
Thanks for your great response . Yes we followed the mentioned document while migrate from ES to Opensearch1.x about 18 months back . Right now its straight forward upgrade to 2.16 which is showing degrade in the performance . Hence not sure what is missing here .
Thanks