Thanks to visit codestin.com
Credit goes to github.com

Skip to content

feat: add gpu monitoring #799

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Open
wants to merge 20 commits into
base: main
Choose a base branch
from
Open

feat: add gpu monitoring #799

wants to merge 20 commits into from
+671 −15

Conversation

@mrtimothyduong
Copy link

@mrtimothyduong mrtimothyduong commented Oct 29, 2025 •
edited by greptile-apps bot
Loading

as per #670, this is an updated PR from main branch, as I dismissed the previous commits to rebase.

Compounded Implementations / PR:

  • folded/merged 'gpuservice.go' into 'system_handler.go' as per feedback
  • updated both dockerfiles with gcompat & added library pathing for nvidia-smi for all different architectures, testing failed when I re-tested, thus adding library pathing.
  • added examples/compose.agent-gpu-nvidia.yaml for example for agent deployment + Intel/AMD/Nvidia. Please feel free to reject or not.
  • added and updated frontend/src/lib/components/gpu-meter.svelte to use bytes package. Feedback incorporated.
  • updated frontend/src/routes/dashboard/+page.svelte dashboard to update based on if GPU agent is sending data or not. Feedback incorporated.

Testing

  • tested frontend on ubuntu docker host (mrtimothyduong/arcane-gpu-mon:1.7.0-gpu)
  • tested agent (headless) on ubuntu docker host with nvidia gpu (mrtimothyduong/arcane-gpu-mon-agent:1.7.0-gpu)
image

Disclaimer Greptiles Reviews uses AI, make sure to check over its work

Greptile Overview

Updated On: 2025-10-30 00:31:15 UTC

Greptile Summary

This review covers only the changes made since the last review, not the entire PR. This PR implements comprehensive GPU monitoring functionality for Arcane, adding real-time GPU statistics display to the dashboard. The implementation supports multiple GPU vendors (NVIDIA, AMD, Intel) through vendor-specific command-line tools and integrates GPU metrics into the existing WebSocket-based system stats.

Key changes include: GPU detection and stats collection logic in system_handler.go with caching mechanisms, a new GpuMeter Svelte component for dashboard display, Docker image updates with gcompat package for GPU tool compatibility, comprehensive multi-architecture library path configuration, and vendor-specific Docker Compose examples for different GPU deployments. The frontend integration conditionally displays GPU metrics when available and follows existing dashboard patterns for consistency.

Important Files Changed
Filename Score Overview
backend/internal/api/system_handler.go 3/5 Adds comprehensive GPU monitoring with vendor detection, stats collection, and caching - complex logic with potential reliability issues
frontend/src/lib/components/gpu-meter.svelte 4/5 New Svelte5 component for GPU statistics display with proper state management and formatting
frontend/src/routes/dashboard/+page.svelte 4/5 Integrates GPU metrics into dashboard with conditional display and historical tracking
docker/Dockerfile 4/5 Adds gcompat package and extensive library paths for GPU tool compatibility across architectures
docker/Dockerfile-agent 4/5 Similar GPU support additions for agent deployment with comprehensive library path configuration
examples/compose.gpu-nvidia.yaml 5/5 Clean Docker Compose example for NVIDIA GPU deployment with proper runtime configuration
examples/compose.gpu-amd.yaml 4/5 AMD GPU deployment example with correct device passthrough and ROCm environment variables
examples/compose.gpu-intel.yaml 4/5 Intel GPU deployment example using DRI device passthrough
examples/compose.agent-gpu-nvidia.yaml 4/5 Agent-specific NVIDIA GPU deployment configuration
frontend/src/lib/types/system-stats.type.ts 5/5 Clean TypeScript interface additions for GPU statistics with proper optional handling
frontend/messages/en.json 5/5 Simple i18n string additions for GPU meter UI text
docker/Dockerfile-agent-static 4/5 Adds gcompat package for GPU tool compatibility in static builds
docker/Dockerfile-static 4/5 Similar gcompat addition for static frontend builds
.gitignore 5/5 Clean exclusion of development-specific files and compose examples

Confidence score: 3/5
  • This PR introduces complex GPU monitoring functionality that may fail on systems without proper GPU drivers or tools installed
  • Score reflects concerns about error handling robustness in GPU detection logic, potential command execution failures, and lack of comprehensive testing across different GPU vendors and configurations
  • Pay close attention to backend/internal/api/system_handler.go for error handling patterns and potential runtime failures when GPU tools are unavailable

Context used:

  • Context from dashboard - .github/copilot-instructions.md file (source)

mrtimothyduong and others added 13 commits October 16, 2025 16:44
@mrtimothyduong
Feat: Add GPU Monitoring for agent & agent-headless
50b3b52
@mrtimothyduong
Merge branch 'main' into main
c517f5f
@mrtimothyduong
feat: updated dashboard to use 4 columns for large screens
565b199
@mrtimothyduong
Merge branch 'main' of https://github.com/mrtimothyduong/arcane-gpu-mon
f7735d9
@mrtimothyduong @greptile-apps
Update frontend/src/lib/components/gpu-meter.svelte
f63af14 
Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com>
@kmendell
Merge branch 'main' into main
6a1605d
@kmendell
fix linter
2d53025
@kmendell
Merge branch 'main' into main
8c9daa3
@mrtimothyduong
feat: add gpu monitoring into system_handler.go
aa2c266
@mrtimothyduong
Merge branch 'main' of https://github.com/mrtimothyduong/arcane-gpu-mon
7193cea
@mrtimothyduong
fix: used 'bytes' to convert binary
57dbda4
@mrtimothyduong
fix: Restored some code from system_handler.go
645dadf 
fix: Removed untested compose.yaml files (AMD/Intel)
@mrtimothyduong
fix: merged gpu_service.go into system_handler.go
a8e7fa5
@mrtimothyduong mrtimothyduong requested a review from a team as a code owner October 29, 2025 11:05
@kmendell
Copy link
Member

kmendell commented Oct 30, 2025

@greptileai

greptile-apps[bot]
greptile-apps bot reviewed Oct 30, 2025
View reviewed changes
Copy link

@greptile-apps greptile-apps bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

14 files reviewed, 10 comments

Edit Code Review Agent Settings | Greptile

@kmendell
Copy link
Member

kmendell commented Oct 30, 2025

Besides the Greptile feedback it looks good to me. Once thats addressed this should be fine i think.

@mrtimothyduong
fix: addressed system_handler bytes conversion
bd92929 
fix: test sh commands for compose.gpu-*.yaml
fix: addressed system_handler.go gpu percentages
fix: removed redundant code from system_handler.go
fix: removed gpu booleans from system_handler.go
@mrtimothyduong
Merge branch 'main' of https://github.com/mrtimothyduong/arcane-gpu-mon
0adbee2
@mrtimothyduong
Merge branch 'main' into main
3be3db0
@mrtimothyduong
Copy link
Author

mrtimothyduong commented Oct 30, 2025 •
edited
Loading

Need to rebuild and test. I have committed for the time being.

Edit:
Pulled from 1.7.1 main/branch. Deployed and tested internally without issues.

@kmendell
Copy link
Member

kmendell commented Oct 31, 2025

Looks good to me the last thing ill ask of you instead of the compose examples here, can we put tjem on the website maybe? either way some docuentation surounding this feature would be nice. I am preapring ot leave for a trip here soon, so i am just a little all over the place https://github.com/ofkm/arcane-website

@kmendell kmendell requested a review from a team as a code owner November 4, 2025 19:09
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

1 more reviewer

@greptile-apps greptile-apps[bot] greptile-apps[bot] left review comments

Reviewers whose approvals may not affect merge requirements

At least 1 approving review is required to merge this pull request.

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@mrtimothyduong @kmendell