Webserver load issues with ES 5.1.2

[barry1]

Hello,

We've been working the last couple of weeks to make our application work with ES 5.1.2 (current version is 2.3.1). After testing everthing, from indexing documents to searching document and everything in between, all seemed fine. So we uploaded the code and switched production to the new 5.1.2 cloud cluster.

But then the load on our webservers, which do the search requests to the elastic cloud cluster, skyrocketed from about 1 to 50. So we reverted all changes and swiched back to the 2.3.1 cluster and everything stabilized again.

The setup is the same on both versions, about 60 indices with 4 shards each. Mapping didn't change exept all string fields have become text fields.

Does anyone have the same issues with ES 5.1.x? Or does anyone have any thoughts on what to change and/or test to solve this issue?

PS. I'm not sure if this belongs in de Elasticsearch or the Elastic Cloud category, feel free to move it if necessary.

PPS. Maybe this is related to this topic, but the cluster itself seems to handle everything just fine.

[scottw (Scott Watermasysk 📎)]

Hi Barry,

I saw a minor improvement when I moved from 5.1.1 to 5.1.2.

The ElasticCloud folks told me last week there was an issue on my cluster itself they were going to fix by Friday. But unfortunately it has been total silence from them since then.

-Scott

[barry1]

Hi @scottw

We tested on 5.1.1 first, then the load on our webservers raised to 200, now with 5.1.2 it raised to 50, but reverted the upload sooner.

We have had some other issues with the new cluster:

Updating scripts configuration doesn't work
Can’t Access Config, Security or Snapshots on Existing Cluster