# Quantitative Cluster Sizing

**URL:** <https://discuss.elastic.co/t/quantitative-cluster-sizing/219850>\
**Category:** Elasticsearch\
**Created:** [February 18, 2020, 6:56pm UTC](https://discuss.elastic.co/t/quantitative-cluster-sizing/219850 "2020-02-18T18:56:45Z")\
**Posts on this page:** 4\
**Page:** 1

<div class="post-metadata">

**Author:** ![sks1234](https://avatars.discourse-cdn.com/v4/letter/s/71c47a/32.png) [@sks1234](https://discuss.elastic.co/u/sks1234)\
**Post date:** [February 18, 2020, 6:56pm UTC](https://discuss.elastic.co/t/quantitative-cluster-sizing/219850/1 "2020-02-18T18:56:45Z")

</div>

Hi I am looking for Sizing Elastic Search Cluster and came across below session which is very informative.  
[https://www.elastic.co/webinars/elasticsearch-sizing-and-capacity-planning](https://www.elastic.co/webinars/elasticsearch-sizing-and-capacity-planning) I have few questions regarding the session -

1. Cluster Sizing for Volume -\> As per session we can have Memory:Storage Ratio of 1:30 for HOT , 1:1000 for Cold etc. so suppose i have 8 gb ram on a cold node then i can store 8000 GB data.
2. Now if we consider number of shards.It is driven by Heap memory. We can have max 4 gb ram for heap so we can support at max 4\* 20 =80 shards per node. One shard can be of max 50 GB so total supported data size = 80 \* 50 = 4000 gb.

So 1st point in providing different storage data than 2nd point. Please help me understand.

---

<div class="post-metadata">

**Author:** ![Christian\_Dahlqvist](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/christian_dahlqvist/32/4617_2.png) [@Christian\_Dahlqvist](https://discuss.elastic.co/u/Christian_Dahlqvist)\
**Post date:** [February 18, 2020, 8:26pm UTC](https://discuss.elastic.co/t/quantitative-cluster-sizing/219850/2 "2020-02-18T20:26:23Z")

</div>

The ratio mentioned for cold nodes assumes the use of frozen indices as frozen indices use very little heap space which allows very dense nodes.

The fact that a maximum of 20 shards per GB of heap is often recommended does however not mean that a node can hold that many shards irrespective of size. If you use very large shards you are likely to have less than 20 per GB of heap in my experience.

---

<div class="post-metadata">

**Author:** ![sks1234](https://avatars.discourse-cdn.com/v4/letter/s/71c47a/32.png) [@sks1234](https://discuss.elastic.co/u/sks1234)\
**Post date:** [February 24, 2020, 5:35am UTC](https://discuss.elastic.co/t/quantitative-cluster-sizing/219850/3 "2020-02-24T05:35:21Z")

</div>

Thanks.

---

<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:** [March 23, 2020, 5:35am UTC](https://discuss.elastic.co/t/quantitative-cluster-sizing/219850/4 "2020-03-23T05:35:26Z")

</div>

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