# APM server recommendations

**URL:** https://discuss.elastic.co/t/apm-server-recommendations/145447
**Category:** APM
**Created:** [August 21, 2018, 6:55pm UTC](https://discuss.elastic.co/t/apm-server-recommendations/145447 "2018-08-21T18:55:04Z")
**Posts on this page:** 5
**Page:** 1

<div class="post-metadata">

### Author: ![kd5dbl](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/kd5dbl/32/10450_2.png) [@kd5dbl](https://discuss.elastic.co/u/kd5dbl)
#### Post date: [August 21, 2018, 6:55pm UTC](https://discuss.elastic.co/t/apm-server-recommendations/145447/1 "2018-08-21T18:55:05Z")

</div>

I'm going to setup an APM server in AWS and was wondering if there were any recommendations on the resource requirements for this server?

I know it's a little vague as I can't give you full details as to how much stuff I'm going to be putting in to it, but was wondering is APM server more memory bound or CPU?

Would an R4 class do? Or do I need to step up to a C4 class box?

Thanks!

--  
David

---

<div class="post-metadata">

### Author: ![jalvz](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/jalvz/32/45613_2.png) [@jalvz](https://discuss.elastic.co/u/jalvz)
#### Post date: [August 22, 2018, 7:52am UTC](https://discuss.elastic.co/t/apm-server-recommendations/145447/2 "2018-08-22T07:52:55Z")

</div>

Hi!

Thanks for trying Elastic APM! As you say it depends a lot on the workload; APM Server usually is memory bound, and for most cases 1Gb RAM is enough. A T3 or something like that could also work.

If you find yourself needing more, keep an eye on Elasticsearch storage too, because is more likely you will hit that limit first.  
Is hard to give exact numbers, but in the documentation you can find some pointers on how to trade off amount of data and cost, for instance here: [https://www.elastic.co/guide/en/apm/server/master/tune-data-ingestion.html](https://www.elastic.co/guide/en/apm/server/master/tune-data-ingestion.html) and here: [https://www.elastic.co/guide/en/apm/server/master/storage-management.html](https://www.elastic.co/guide/en/apm/server/master/storage-management.html)

I hope this helps, will be happy to hear how is your experience!

---

<div class="post-metadata">

### Author: ![kd5dbl](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/kd5dbl/32/10450_2.png) [@kd5dbl](https://discuss.elastic.co/u/kd5dbl)
#### Post date: [August 22, 2018, 12:00pm UTC](https://discuss.elastic.co/t/apm-server-recommendations/145447/3 "2018-08-22T12:00:29Z")

</div>

OK... since we're still just starting to look, I'll start off with something small... R4.large or M4.large maybe....

We've been bitten by the CPU credits with the T2/T3 class before...

I'll try to post something once we have it up and running...

Thanks!

--  
David

---

<div class="post-metadata">

### Author: ![roncohen](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/roncohen/32/22881_2.png) [@roncohen](https://discuss.elastic.co/u/roncohen)
#### Post date: [September 4, 2018, 12:51pm UTC](https://discuss.elastic.co/t/apm-server-recommendations/145447/5 "2018-09-04T12:51:37Z")

</div>

Hi Martha,

I wouldn't recommend running APM Server in a production scenario with less than 500mb of available memory at this point - and even then, you should adjust some of the default settings in order to run under those conditions. For example, make `queue.mem.events` smaller.

Future versions of APM Server will improve upon this.

best,  
Ron

---

<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: [September 25, 2018, 10:11am UTC](https://discuss.elastic.co/t/apm-server-recommendations/145447/7 "2018-09-25T10:11:45Z")

</div>

This topic was automatically closed 20 days after the last reply. New replies are no longer allowed.
