No CPU Usage data in Proxmox LXC

[Drallas]

Summary

I moved Cronicle over to a LXC container, everything is great except the absence of CPU usage data, memory data is recorded.

Steps to reproduce the problem

Run Cronicle in a tteck LXC.

Your Setup

Proxmox and LXC.

Operating system and version?

Debian 12

Node.js version?

Latest

Cronicle software version?

Latest

Are you using a multi-server setup, or just a single server?

Single server.

Are you using the filesystem as back-end storage, or S3/Couchbase?

B2

[mikeTWC1984]

CPU/memory metrics come from this command

/bin/ps -eo "ppid pid %cpu rss"

You probably have some in-the-box / non-standard version of ps (might need to install standard version)

[Drallas]

CPU/memory metrics come from this command

/bin/ps -eo "ppid pid %cpu rss"

You probably have some in-the-box / non-standard version of ps (might need to install standard version)

Not sure, if I paste in the command I do get output.

/bin/ps -eo "ppid pid %cpu rss"
PPID PID %CPU RSS
0 1 0.0 7808
1 51 0.0 7648
1 73 0.0 3456
1 91 0.0 2176
1 100 0.0 1408
1 101 0.0 2944
1 104 0.0 5248
1 156 0.0 1152
1 158 0.0 1024
1 159 0.0 1152
1 160 0.0 3968
1 167 0.2 52612
1 342 0.0 2964
160 13543 0.0 4352
1 13546 0.0 6144
13546 13547 0.0 2700
13543 13563 0.0 1536
13563 16737 0.0 1792
16737 16738 0.0 1060
16738 16739 0.0 1280
342 453070 0.0 6912
342 453071 0.0 6784
160 471751 0.0 11264
471751 471755 0.0 8960
471755 478254 0.0 4096

Seems my jobs are too short, to display average CPU time, only some heavier ones have a value now.

[mikeTWC1984]

Oh, ok. TO test, you can create event using test plugin, and check "burn cpu" option. Normally it should utilize 1 core for 100%. Also, performance metrics get update after 10 seconds

[Drallas]

Most of my jobs are between 5 and 20 seconds. Since it an LXC on Proxmox the CPU impact might be to low to get over the threshold to be picked up. Only one job that runs longer shows CPU stats.