# Investigating high heap usage

**URL:** <https://discuss.elastic.co/t/investigating-high-heap-usage/62574>\
**Category:** Elasticsearch\
**Created:** [October 9, 2016, 9:28am UTC](https://discuss.elastic.co/t/investigating-high-heap-usage/62574 "2016-10-09T09:28:13Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![orweinberger](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/orweinberger/32/580_2.png) [@orweinberger](https://discuss.elastic.co/u/orweinberger)\
**Post date:** [October 9, 2016, 9:28am UTC](https://discuss.elastic.co/t/investigating-high-heap-usage/62574/1 "2016-10-09T09:28:13Z")

</div>

What is the best approach or way to investigate why heap usage is so high? I'm watching the node stats and I can see the fielddata and filter cache which I understand are a part of the heap, but if those two are low, no indexing is currently being done and heap is still high, where should I be looking?

Another thing I tried is getting the `hot_threads`, but I'm not sure what to make of it:

> <https://gist.github.com/orweinberger/3853094d8b204e481b70b861ffbdf5e6>

Is there a way to get a snapshot of whats currently in the heap in a human readable way?

---

<div class="post-metadata">

**Author:** ![warkolm](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/warkolm/32/39224_2.png) [@warkolm](https://discuss.elastic.co/u/warkolm)\
**Post date:** [October 10, 2016, 8:10am UTC](https://discuss.elastic.co/t/investigating-high-heap-usage/62574/2 "2016-10-10T08:10:18Z")

</div>

Are you using Marvel to monitor things?

> [@orweinberger](#):
>
> Is there a way to get a snapshot of whats currently in the heap in a human readable way?

Nope ☹

---

<div class="post-metadata">

**Author:** ![orweinberger](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/orweinberger/32/580_2.png) [@orweinberger](https://discuss.elastic.co/u/orweinberger)\
**Post date:** [October 10, 2016, 12:21pm UTC](https://discuss.elastic.co/t/investigating-high-heap-usage/62574/3 "2016-10-10T12:21:57Z")

</div>

Unfortunately I'm not using Marvel, is there sort of a cheap sheet of things to check in order to diagnose heap issuse in Elasticsearch? Is there a page that explains how to read the `hot_threads` output?

---

<div class="post-metadata">

**Author:** ![warkolm](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/warkolm/32/39224_2.png) [@warkolm](https://discuss.elastic.co/u/warkolm)\
**Post date:** [October 10, 2016, 11:12pm UTC](https://discuss.elastic.co/t/investigating-high-heap-usage/62574/4 "2016-10-10T23:12:17Z")

</div>

No and no ☹  
But a quick check doesn't show anything bad in that hot threads.

---

<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 5, 2017, 10:13pm UTC](https://discuss.elastic.co/t/investigating-high-heap-usage/62574/5 "2017-07-05T22:13:29Z")

</div>