[Deployment] System Monitoring

[KY00KIM]

Test deployment & monitoring (2 Group)

Test setup :

- 1 Instance
    - OS : Ubuntu20.04
    - CPU : Intel 8255C 2.5Ghz 8core
    - MEM : 32GB
    - GPU : NVIDIA T4(16GB)
- Deployment
    1. api-server1 + webui
    2. api-server2 + webui

graph LR;
subgraph Instance-T4
subgraph Group1
            api-server1 --> webui1;
end
subgraph Group2
            api-server2--> webui2;
end

end
    RabbitMQ --> api-server1;
    RabbitMQ --> api-server2;

Max memory usage

- on init :
    - MEM : 5.5GB
    - VRAM : 5GB
- on simultaneous 2 request
    - MEM : 18.8GB
    - VRAM : 14.9GB

[KY00KIM]

Test deployment & monitoring (1 Group)

Test setup :

- 1 Instance
    - OS : Ubuntu20.04
    - CPU : Intel 8255C 2.5Ghz 8core
    - MEM : 32GB
    - GPU : NVIDIA T4(16GB)
- Deployment
    1. api-server1 + webui

graph LR;
subgraph Instance-T4
subgraph Group1
            api-server1 --> webui1;
end

end
    RabbitMQ --> api-server1;

Max memory usage

- on init (2024.04.28):
    - MEM : 8GB
    - VRAM : 7GB
- 2일 경과 후(2024.04.30)
    - MEM : 18.8GB
    - VRAM : 14.9GB
    - Process 별 자원 점유 현황
      - api-server : CPU 0.5%  / MEM 7.1%(2.2GB)
      - webui : CPU 4.5%  / MEM 56.2%(18GB)

[seooyxx]

https://github.com/AUTOMATIC1111/stable-diffusion-webui/discussions/6722

이 이슈 참고해 보면 좋을 것 같습니다

[KY00KIM]

TCMalloc / Docker image build 수정 확인이요