# Metricbeat metrics collection process and period/aggregation

**URL:** <https://discuss.elastic.co/t/metricbeat-metrics-collection-process-and-period-aggregation/90855>\
**Category:** Beats\
**Tags:** metricbeat\
**Created:** [June 26, 2017, 4:43pm UTC](https://discuss.elastic.co/t/metricbeat-metrics-collection-process-and-period-aggregation/90855 "2017-06-26T16:43:50Z")\
**Posts on this page:** 4\
**Page:** 1

<div class="post-metadata">

**Author:** ![ELK\_GB](https://avatars.discourse-cdn.com/v4/letter/e/a9adbd/32.png) [@ELK\_GB](https://discuss.elastic.co/u/ELK_GB)\
**Post date:** [June 26, 2017, 4:43pm UTC](https://discuss.elastic.co/t/metricbeat-metrics-collection-process-and-period-aggregation/90855/1 "2017-06-26T16:43:50Z")

</div>

Hi All,

I'm trying to understand something about the metricbeat metrics collection mechanism.

My question is if I leave my metricbeat configs at defaults i.e. period set to 10s and outputting to local file, ES or LS, how does metricbeat work when it runs, does it look at history data and take the average of the metric or does it actually get a single value right when its time to run the metricsets ?

for example if I look at the raw JSON output of metricbeat are the CPU metrics values i'm seeing for one CPU event based on a specific criteria in the metricbeat design i.e. does it aggregate in the background or uses certain criteria to determine those values.

Thanks

---

<div class="post-metadata">

**Author:** ![steffens](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/steffens/32/79630_2.png) [@steffens](https://discuss.elastic.co/u/steffens)\
**Post date:** [June 27, 2017, 8:27am UTC](https://discuss.elastic.co/t/metricbeat-metrics-collection-process-and-period-aggregation/90855/2 "2017-06-27T08:27:31Z")

</div>

Metricbeat is collecting and shipping the raw values, as reported by the OS/services it monitors. Right when the metricset is executed. (Almost) No additional aggregations/computations do take place. Aggregations/rates are the computed by Elasticsearch/Kibana on query time.

CPU usage is a little special, as CPU usage us reported in ticks and need to be normalized to a percentage + one needs to know the number of cores. But besides some minimal normalization if required, no additional aggregations are done within metricbeat.

---

<div class="post-metadata">

**Author:** ![ELK\_GB](https://avatars.discourse-cdn.com/v4/letter/e/a9adbd/32.png) [@ELK\_GB](https://discuss.elastic.co/u/ELK_GB)\
**Post date:** [July 5, 2017, 9:21pm UTC](https://discuss.elastic.co/t/metricbeat-metrics-collection-process-and-period-aggregation/90855/3 "2017-07-05T21:21:04Z")

</div>

@steffens Thank you! I appreciate your response.

---

<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:** [August 2, 2017, 9:21pm UTC](https://discuss.elastic.co/t/metricbeat-metrics-collection-process-and-period-aggregation/90855/4 "2017-08-02T21:21:14Z")

</div>

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