# Shard best practices count vs stats for limit

**URL:** <https://discuss.elastic.co/t/shard-best-practices-count-vs-stats-for-limit/375377>\
**Category:** Elasticsearch\
**Created:** [March 4, 2025, 1:45pm UTC](https://discuss.elastic.co/t/shard-best-practices-count-vs-stats-for-limit/375377 "2025-03-04T13:45:37Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![zedEm](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/zedem/32/141771_2.png) [@zedEm](https://discuss.elastic.co/u/zedEm)\
**Post date:** [March 4, 2025, 1:45pm UTC](https://discuss.elastic.co/t/shard-best-practices-count-vs-stats-for-limit/375377/1 "2025-03-04T13:45:37Z")

</div>

I have a question about ES best practices as written here [Size your shards | Elasticsearch Guide [8.17] | Elastic](https://www.elastic.co/guide/en/elasticsearch/reference/current/size-your-shards.html).

During inspections I have come across index size discrepancy with stats vs count. This is most likely due to having nested objects.([the maximum number of documents in an elasticsearch index and difference about stats and count](https://stackoverflow.com/questions/78611926/the-maximum-number-of-documents-in-an-elasticsearch-index-and-difference-about-s))

So, when sizing shards as written in the documents above, it is a good practice to go up from 10 to 50Gb and document count up to 200M.

For the 200M count, should I take a stats or count number?