# Elasticsearch \_stats api metric questions

**URL:** <https://discuss.elastic.co/t/elasticsearch-stats-api-metric-questions/211429>\
**Category:** Elasticsearch\
**Created:** [December 11, 2019, 8:08am UTC](https://discuss.elastic.co/t/elasticsearch-stats-api-metric-questions/211429 "2019-12-11T08:08:49Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![sujieim](https://avatars.discourse-cdn.com/v4/letter/s/2bfe46/32.png) [@sujieim](https://discuss.elastic.co/u/sujieim)\
**Post date:** [December 11, 2019, 8:08am UTC](https://discuss.elastic.co/t/elasticsearch-stats-api-metric-questions/211429/1 "2019-12-11T08:08:49Z")

</div>

Hi,  
I am using \_stats API for collecting cluster metrics.  
I cannot find enough information on details of each metrics, and so having trouble to use those on a monitoring server.

1. I assumed certain fields are accumulative (e.g. query\_total, query\_time\_in\_millis) and didn't concern about those values decreasing. As I was applying derivatives to those fields, I was getting minus values and couldn't interpret further. In which case does query\_time\_millis reset? or decrease?

2. Is there an alternative field to calculate request latencies?  
I was using query\_total divided by query\_time\_in\_millis, but couldn't resolve the case when query\_time\_in\_millis decreases.

3. What does merge\_docs indicate?

If there is a document or reference explaining details on each stat metrics, it would be a great help.

---

<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:** [January 8, 2020, 8:08am UTC](https://discuss.elastic.co/t/elasticsearch-stats-api-metric-questions/211429/2 "2020-01-08T08:08:57Z")

</div>

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