Help understanding best practice of monitoring Linux guest memory on proxmox

surfrock66

Active Member
Feb 10, 2020
30
7
28
40
I'm using a completely up to date proxmox install and monitoring my Ubuntu VM's using an up to date zabbix. I have noticed a large discrepancy in the memory usage reported in proxmox vs. the memory usage reported by the vm. This makes sense; from the hypervisor side, this is a ballooning device, so if I needed 4GB at some narrow point (like during backups), proxmox is gonna show 4GB of allocation and only request a release from the VM when the host runs out of memory. The issue is, high memory usage is an alarm condition in the proxmox template. I am getting a constant list of high memory problems when the real use of the server is ultimately below 20%, and the problems don't go away because proxmox doesn't request to release the memory until the hypervisor needs it.

What would best practice be here? There's a case for just disabling the trigger on the proxmox template, but at the same time, I am not a huge fan in general of disabling triggers. Is there a way to instruct proxmox to be more aggressive in requesting to reclaim unused memory, or is that just a bad idea in general?

1687970778809.png

1687970836707.png
 
What would best practice be here? There's a case for just disabling the trigger on the proxmox template, but at the same time, I am not a huge fan in general of disabling triggers. Is there a way to instruct proxmox to be more aggressive in requesting to reclaim unused memory, or is that just a bad idea in general?
By default neither ballooning nor KSM will do anything as long as the hosts RAM is below 80% usage.
I monitor my VMs from within the VM with the zabbix agent, so it is showing the VMs available and not the free memory, so I can ignore what PVE is reporting as RAM usage for those VMs.
 
Yea I'm monitoring both, but the question really is should I just disable the trigger prototype in the template. I think that's the move. Thanks!
 
I personally just disabled all monitoring of VMs by making use of the macros ("{$PVE_FORCE_ID}" but I'm not using the official template). I don't need to log and monitor the same VMs twice and I prefer the reports from the guestOS side.
 
Last edited:

About

The Proxmox community has been around for many years and offers help and support for Proxmox VE, Proxmox Backup Server, and Proxmox Mail Gateway.
We think our community is one of the best thanks to people like you!

Get your subscription!

The Proxmox team works very hard to make sure you are running the best software and getting stable updates and security enhancements, as well as quick enterprise support. Tens of thousands of happy customers have a Proxmox subscription. Get yours easily in our online shop.

Buy now!