# Elasticsearch official benchmarking results

**URL:** https://discuss.elastic.co/t/elasticsearch-official-benchmarking-results/156315
**Category:** Elasticsearch
**Tags:** rally
**Created:** [November 12, 2018, 3:55pm UTC](https://discuss.elastic.co/t/elasticsearch-official-benchmarking-results/156315 "2018-11-12T15:55:02Z")
**Posts on this page:** 3
**Page:** 1

<div class="post-metadata">

### Author: ![rajisankar](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/rajisankar/32/43250_2.png) [@rajisankar](https://discuss.elastic.co/u/rajisankar)
#### Post date: [November 12, 2018, 3:55pm UTC](https://discuss.elastic.co/t/elasticsearch-official-benchmarking-results/156315/1 "2018-11-12T15:55:02Z")

</div>

Hi,

I was going through the official Elasticsearch benchmarking results provided online at  
[https://elasticsearch-benchmarks.elastic.co/index.html#tracks/http-logs/nightly/30d](https://elasticsearch-benchmarks.elastic.co/index.html#tracks/http-logs/nightly/30d)

I noticed something intriguing. The numbers using a 3-node Elasticsearch is smaller than the number obtained using the default one node cluster. Why is that? Shouldn't a 3-node cluster provide better indexing numbers that a single node with default configuration?

Am i missing something here?

---

<div class="post-metadata">

### Author: ![dliappis](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/dliappis/32/56174_2.png) [@dliappis](https://discuss.elastic.co/u/dliappis)
#### Post date: [November 13, 2018, 5:36pm UTC](https://discuss.elastic.co/t/elasticsearch-official-benchmarking-results/156315/2 "2018-11-13T17:36:26Z")

</div>

Hi,

Thanks for raising this.

Broadly speaking, yes, 3-node should provide higher indexing throughput.  
Indeed this is the case with other tracks: [nyc taxis](https://elasticsearch-benchmarks.elastic.co/#tracks/nyc-taxis/nightly/90d), [pmc](https://elasticsearch-benchmarks.elastic.co/#tracks/pmc/nightly/90d).

However, one difference between the 3-node and the default benchmarks is that the former has 1 replica, which reduces the initial indexing performance (see also [here](https://www.elastic.co/guide/en/elasticsearch/reference/6.4/tune-for-indexing-speed.html#_disable_refresh_and_replicas_for_initial_loads)). Additionally [http\_log](https://github.com/elastic/rally-tracks/tree/master/http_logs#example-document), as well as [geonames](https://github.com/elastic/rally-tracks/tree/master/geonames#example-document) and [geopoint](https://github.com/elastic/rally-tracks/tree/master/geopoint#example-document) have a very small doc size and get impacted more.

---

<div class="post-metadata">

### Author: ![system](https://us1.discourse-cdn.com/elastic/original/3X/1/a/1ac57faf039f6b580b3f104ef42a2a89e41014de.png) [@system](https://discuss.elastic.co/u/system)
#### Post date: [December 11, 2018, 5:36pm UTC](https://discuss.elastic.co/t/elasticsearch-official-benchmarking-results/156315/3 "2018-12-11T17:36:27Z")

</div>

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