True memory and CPU usage not matching docker stats

Post Reply
User avatar
Qmann
Easy as a breeze
Posts: 302
Joined: Mon Jun 08, 2020 8:09 am
Location: USA

True memory and CPU usage not matching docker stats

Post by Qmann »

Does anyone know how to get the real CPU and memory usage of a container? When it first appeared that memory was maxing, I passed passed the "-e CRASHPLAN_SRV_MAX_MEM=40960M" in creating the container as I wanted to give it nearly everything on memory to see where it topped out. But I cannot tell what it's actually doing. Here is what I see for Crashplan pro as an example during a backup:

Container station data 1:
CS1.png
Container station data 2:
CS2.png
A pure "docker stats" in the CLI on the core NAS:
CS3.png
And finally, and "top" command from inside the docker shell:
CS4.png
Container settings:
CS5.png


What is real? It looks to me the 21% CPU is real as I also did a TOP on the NAS CLI, and it matched the container TOP cpu usage. Resource panel on the NAS shows memory is 89% free which seems real.

So it seems the "RAM at 91%" in the Crashplan container station view is total BS? Or is there some limit on a container I don't know about and that's why it shows 91%?

Q
You do not have the required permissions to view the files attached to this post.
Model: TVS-872XT 64GB (Crucial 64GB Kit CT2K32G4SFD8266)
2 x 1TB XPG 1TB NVMe (ASX8200PNP-1TT-C) [RAID-1]
5 x 16TB EXOS [RAID-5]
Borg Backup running to an offisite pi, AND to the local TS-569L
Model: TS-569L Borg server for backups
6 x 8TB Ironwolf [RAID-5]
Qotom-Q355G4 Fanless Mini Micro PC running pSense in front of everything
haproxy for anything inside the LAN
User avatar
Qmann
Easy as a breeze
Posts: 302
Joined: Mon Jun 08, 2020 8:09 am
Location: USA

Re: True memory and CPU usage not matching docker stats

Post by Qmann »

Any takers?
Model: TVS-872XT 64GB (Crucial 64GB Kit CT2K32G4SFD8266)
2 x 1TB XPG 1TB NVMe (ASX8200PNP-1TT-C) [RAID-1]
5 x 16TB EXOS [RAID-5]
Borg Backup running to an offisite pi, AND to the local TS-569L
Model: TS-569L Borg server for backups
6 x 8TB Ironwolf [RAID-5]
Qotom-Q355G4 Fanless Mini Micro PC running pSense in front of everything
haproxy for anything inside the LAN
datamonkey
New here
Posts: 3
Joined: Tue Sep 29, 2020 5:14 am

Re: True memory and CPU usage not matching docker stats

Post by datamonkey »

I have a somewhat similar question and just opened a new topic on it. My Container Station Overview is showing 74% RAM usage, but the system is only showing 13%. There are 8 LXC containers running, but seven of them show 0% RAM usage and the last one shows 1%. In other words, they aren't using much. I'm not sure why there's a difference between what Container Station is showing and what the system is showing.
Model: TS-h1277XU-RP
FW: h4.5.1.1491
User avatar
Qmann
Easy as a breeze
Posts: 302
Joined: Mon Jun 08, 2020 8:09 am
Location: USA

Re: True memory and CPU usage not matching docker stats

Post by Qmann »

I haven't been able to figure it out and assume that container station is just plain wrong.
Model: TVS-872XT 64GB (Crucial 64GB Kit CT2K32G4SFD8266)
2 x 1TB XPG 1TB NVMe (ASX8200PNP-1TT-C) [RAID-1]
5 x 16TB EXOS [RAID-5]
Borg Backup running to an offisite pi, AND to the local TS-569L
Model: TS-569L Borg server for backups
6 x 8TB Ironwolf [RAID-5]
Qotom-Q355G4 Fanless Mini Micro PC running pSense in front of everything
haproxy for anything inside the LAN
DaveM
Starting out
Posts: 28
Joined: Tue Jul 08, 2008 1:31 am

Re: True memory and CPU usage not matching docker stats

Post by DaveM »

I've also noticed that the containers I have on "host" networking always show a consistent volume of network traffic and exactly the same even though they're all doing nothing. It's obvious the data is coming from higher up the chain I guess.
kovacsgergely83
First post
Posts: 1
Joined: Fri Mar 19, 2021 3:57 pm

Re: True memory and CPU usage not matching docker stats

Post by kovacsgergely83 »

Might be that the Container uses dynamic memory, and can expand from 0 to 8GB (in my case with an memory-upgraded a TS-431P2), but as long as it only requires let's say 700 MB memory at the moment, then is reserves only 1GB for the Conatiner itself, and then the 70% of 1GB is what we can see in Container Station, while in Resource Explorer it is much less?
User avatar
ve7tcc
Starting out
Posts: 17
Joined: Mon Sep 06, 2021 1:32 am

Re: True memory and CPU usage not matching docker stats

Post by ve7tcc »

I have the same problem. I removed all containers, and it says 81%
I stopped QuMaggie, Qsirch, QNAP AI Core, and still 81%
Resource monitor shows only 28% usage.
Awaiting a call from QNAP support.
Model: TS-h973AX
FW: h4.5.4.1800
lexius
Starting out
Posts: 43
Joined: Sat Nov 06, 2010 10:11 pm

Re: True memory and CPU usage not matching docker stats

Post by lexius »

Hello. Did you get an answer from QNAP Support?
TS-H886
Post Reply

Return to “Container Station”