Memory usage indication for OpenVZ containers

[lucaferr]

Hi all!
First of all, congratulations to the developers of this AWESOME virtualization platform!
Second, I have an idea to suggest to the developers: at the moment the Proxmox GUI shows just the "used memory" of a VPS, which includes memory used by the programs + buffers + cache. I think that admins would be interested in knowing how much of this memory is used by applications and how much by buffers + cache. This would make it easier for admins to understand whether the memory allocation is correct or should be raised.

For example, I have a fileserver VPS running Samba and Proxmox shows:

Total: 4.00GB
Used: 3.03GB

I was thinking about raising the dedicated memory, which would have meant to buy more physical RAM...Then inside the VPS I run "vmstat -s -S M | grep mem", obtaining the following output:

4000 M total memory
3087 M used memory
349 M active memory
2675 M inactive memory
912 M free memory
0 M buffer memory

So actually just 349MB is currently used by applications, the rest is cache+buffers...so the good news is that I don't have to buy new RAM!
I would have figured it out immediately if the memory usage panel in the Proxmox GUI was something like:

Total: 4.00GB
Used: 3.03GB --> Applications: 0.35GB

--> Cache + Buffers: 2.68GB
​

Do you agree?
Thank you very much, have a nice day!

 

[lucaferr]

Really, no users are interested in this topic? I now have Proxmox VE 3.4 and it's the same as 2 years ago when I wrote the topic: I have some containers showing 95% memory usage which in reality use less than 50% but I need to SSH-ing into them to figure it out...also graphs in "Summary" GUI is wrong because it makes me think that the VPS is near saturation even though is veeeery far from it...

 

[1nerdyguy]

Really, no users are interested in this topic? I now have Proxmox VE 3.4 and it's the same as 2 years ago when I wrote the topic: I have some containers showing 95% memory usage which in reality use less than 50% but I need to SSH-ing into them to figure it out...also graphs in "Summary" GUI is wrong because it makes me think that the VPS is near saturation even though is veeeery far from it...

This has been the way it is forever, and I see tons of people confused by it. Linux memory management is weird. I get around it by having a monitoring solution, Nagios in my case, watch my machines and report the usage.

Changing the way it is in the Proxmox screen could be a negative thing. I want to know how much is cached and active, total, so I don't oversubscribe my RAM.

 

[lucaferr]

I understand your point. I'm not saying that the total memory should be removed, but only that an indication about the active memory should be added, so every sysadmin has all the informations to understand the usage of the nodes. So instead of just displaying "Memory: 98%", a useful indication would be "Memory: 98% (56% active)"