Excessive usage of cpu

[meower1]

Hi, so i've setup the project on my private server with the following specs

OS: Ubuntu 22.04.4 LTS x86_64
CPU: AMD Ryzen 9 7950X3D (2) @ 4.191GHz
Memory: 2133MiB / 5890MiB ddr5

and snidust seems to be consuming an unreasonable amount of resources specially cpu. it happens after a period of time has passed which im unsure of.

heres the output of docker stats

8b55bbbe447b   sniDust            160.70%   14.34MiB / 5.752GiB   0.24%     22.2GB / 22.1GB   28.6MB / 21.7MB   25

there is currently no user on the dns server other than myself

output of docker logs snidust

[INFO] [SniDust] Adding domain yallo.tv to list
[INFO] [SniDust] *** End of Domain List ***
[INFO] [SniDust] *** Complete! ***
Added downstream server 8.8.8.8:853
Added downstream server 8.8.4.4:853
Added downstream server 1.1.1.1:443
Added downstream server 1.0.0.1:443
Listening on 0.0.0.0:5300
ACL allowing queries from: 0.0.0.0/0
Console ACL allowing connections from: 127.0.0.0/8, ::1/128
Marking downstream cloudflare-dns (1.0.0.1:443) as 'up'
Marking downstream cloudflare-dns (1.1.1.1:443) as 'up'
Marking downstream dns.google (8.8.4.4:853) as 'up'
Marking downstream dns.google (8.8.8.8:853) as 'up'
===================================================================
[INFO] SniDust started => Using server_ip - Point your DNS settings to this address
===================================================================

my docker compose file:

version: '3.8'
services:
    seji64:
        container_name: sniDust
        environment:
            - 'ALLOWED_CLIENTS=0.0.0.0/0'
            - 'EXTERNAL_IP=server_ip'
            - 'SPOOF_ALL_DOMAINS=true' # Set to true (case sensitive!) if you want to spoof ALL domains.

            - TZ=Europe/Berlin
        ports:
            - '443:8443'
            - '80:8080'
            - '53:5300/udp'
            - '53:5300/tcp'
        image: 'ghcr.io/seji64/snidust:main'
        restart: on-failure:5

this has been the case for a while and had happened to me a bunch of time with this project i didn't have the time to report it tho.

thanks for the amazing project tho hoping to see this fixed in the future <3

[Seji64]

could you run htop or something to track down the procress?

[meower1]

this is htop's results filtered to "snidust". the cpu pressure hasn't began yet as i've just started the process and it often starts a few hours after execution

[Seji64]

okay sorry, instructions were a bit unclear....i meant to run htop or similar inside container when the cpu usage is that high

[meower1]

yeah i'll try to replicate the behavior on a spare server

[meower1]

this is the top results from within the container after about a week of having the container running

output of docker stats

and this is the top results from the first day i ran the container

output of `htop` outside of the container on the host machine

so in short the first few days i've ran it i've had no issues with it whatsoever and it ran normally consuming a reasonable amount of resources and it gradually increases the resource usage after a while. restarting the container temporarily solves the issue

[Seji64]

so it seems to be a nginx thing. i applied some changes....(no clue if it helps or make things worse to be honest)

However, i guess if my general use config is not working / scaling for your amout of traffic, you should link your own which optimized for your hardware

[meower1]

no worries ill test the new update and post the results here.

you should link your own which optimized for your hardware

im not sure i understand could u clarify this

[Seji64]

i mean for example you need to set the nginx workers from auto to 16

In this case you can just link your own optimized nginx config:

version: '3.3'
services:
    snidust:
        container_name: snidust
        environment:
            - 'ALLOWED_CLIENTS=127.0.0.1, 10.111.123.7'
            - EXTERNAL_IP=10.111.123.8
        ports:
            - '443:8443'
            - '80:8080'
            - '53:5300/udp'
        volumes:
            - '~/mysuperchargednginx.conf:/etc/nginx/nginx.conf:ro
        image: 'ghcr.io/seji64/snidust:1.0.14'