# High cpu on fresh empty server

**URL:** https://forums.suse.com/t/high-cpu-on-fresh-empty-server/17986
**Category:** SUSE Rancher Prime
**Created:** [July 19, 2020, 2:07am UTC](https://forums.suse.com/t/high-cpu-on-fresh-empty-server/17986 "2020-07-19T02:07:25Z")
**Posts on this page:** 3
**Page:** 1

<div class="post-metadata">

### Author: ![binary64](https://sea2.discourse-cdn.com/flex022/user_avatar/forums.suse.com/binary64/32/2206_2.png) [@binary64](https://forums.suse.com/u/binary64)
#### Post date: [July 19, 2020, 2:07am UTC](https://forums.suse.com/t/high-cpu-on-fresh-empty-server/17986/1 "2020-07-19T02:07:25Z")

</div>

Hi, I set up latest rancher on a single-node digitalocean VPS (1 cpu, 2 gb ram, 8 gb swap), and configured a password, no workloads… just left it running for a week, and the cpu graph looks like

 ![image](https://us1.discourse-cdn.com/flex022/uploads/suse/original/2X/5/5f391895c93531656b9557ea321be605d5176d7a.png)

Have I done something wrong, is this normal, how can I trim it down to like under 5%?

I know you recommend 8 gb ram for nodes and 2 separate ones, but I can’t figure what it’s doing at 30% constantly

Many thanks in advance

---

<div class="post-metadata">

### Author: ![binary64](https://sea2.discourse-cdn.com/flex022/user_avatar/forums.suse.com/binary64/32/2206_2.png) [@binary64](https://forums.suse.com/u/binary64)
#### Post date: [July 19, 2020, 2:09am UTC](https://forums.suse.com/t/high-cpu-on-fresh-empty-server/17986/2 "2020-07-19T02:09:00Z")

</div>

![image](https://us1.discourse-cdn.com/flex022/uploads/suse/original/2X/7/7f109d332def7f6194a3c90bbaabb291d95fca36.png)

---

<div class="post-metadata">

### Author: ![vincent](https://sea2.discourse-cdn.com/flex022/user_avatar/forums.suse.com/vincent/32/7156_2.png) [@vincent](https://forums.suse.com/u/vincent)
#### Post date: [July 19, 2020, 4:16am UTC](https://forums.suse.com/t/high-cpu-on-fresh-empty-server/17986/3 "2020-07-19T04:16:20Z")

</div>

Rancher only actually runs in a kubernetes cluster; the single container image is really running an entire cluster inside of it. K8s/its controllers are what they are, and a lot of what they are is relatively tight loops watching for desired vs actual state differences.