# Detect anomalies for specified time

**URL:** <https://discuss.elastic.co/t/detect-anomalies-for-specified-time/187874>\
**Category:** Elasticsearch\
**Tags:** elastic-stack-machine-learning\
**Created:** [June 27, 2019, 3:28pm UTC](https://discuss.elastic.co/t/detect-anomalies-for-specified-time/187874 "2019-06-27T15:28:18Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![Bharath\_Kumar\_R](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/bharath_kumar_r/32/45474_2.png) [@Bharath\_Kumar\_R](https://discuss.elastic.co/u/Bharath_Kumar_R)\
**Post date:** [June 27, 2019, 3:28pm UTC](https://discuss.elastic.co/t/detect-anomalies-for-specified-time/187874/1 "2019-06-27T15:28:18Z")

</div>

![26%20PM](https://us1.discourse-cdn.com/elastic/original/3X/e/7/e7a7912cfa9e438a76f7c79a6bd2ce3a53f8ea0c.png)  
From the image, we can see that the job has triggered 3 anomalies. 1 severe and 2 minor anomalies.

I am ok with the 1st anomaly (minor) since it Is just a temporary drop.

I am also ok with the 2nd anomaly (critical) since there was a huge drop

However I'm not ok with the 3rd anomaly (minor) since the drop went to 0 and it stayed 0 for a long period of time. Also, according to me this is the most critical of all 3 since no data entered into my system. How to tune parameters so that these kind of anomalies are given the highest weightage

---

<div class="post-metadata">

**Author:** ![richcollier](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/richcollier/32/115035_2.png) [@richcollier](https://discuss.elastic.co/u/richcollier)\
**Post date:** [June 28, 2019, 1:30pm UTC](https://discuss.elastic.co/t/detect-anomalies-for-specified-time/187874/2 "2019-06-28T13:30:58Z")

</div>

I disagree - the 3rd anomaly also contained critical scores (can see red, multi-bucket anomalies). Also, although the leading edge of the deep trough is a yellow (minor) anomaly now, it could have been scored higher at the initial moment of it occurring. You could check the `intitial_record_score` of that anomaly to see what the value was at the time of it being created.

---

<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:** [July 26, 2019, 1:31pm UTC](https://discuss.elastic.co/t/detect-anomalies-for-specified-time/187874/3 "2019-07-26T13:31:08Z")

</div>

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