Load is not distributed across the nodes


We are using elasticsearch 7.6.2. We see that load is not distributed across the nodes.

For anyone to be able to help I think you will need to provide a lot more information about your cluster, configuration and use case as well as details about how the load is unevenly distributed.

Hi Christian,

We have 8 node cluster, 2 client & 6 masterdata nodes.
CPU - 4 each server
RAM -32 each Server.
Storage - 3 TB (6 masterdata)

Business is using python script to upload data into elasticsearch.

Is the python script distributing requests across all data nodes or just sending it to one? What type of load is the cluster under? Are indices and shards evenly distributed?

Hi Christian,

Basically we are using different indices with 1 shard & 1 Replica, We see that CPU usage is more 2 nodes.
In script business is using all masterdata nodes.

Currently we have 344 indices


Is load perhaps higher on indices located on these nodes? Do you have any indices that receive more requests than others?

How to check that.

Do you have monitoring installed? Can you track it from the application?

yes i have installed

How distribute primary & replica equal between the nodes.
I see 1st node has primary and 2nd node has replica only.

Do you insert new data or update existing documents? What does the workload and use case look like?

New indices as well as update of indices. Basically log pertaining to ticket

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.