Analysis updated 2026-05-18
Get a Feishu alert whenever a teammate starts or finishes a GPU job on a shared server.
Track GPU utilization, memory, temperature, and power history without a full monitoring stack.
Run automatic, no-root GPU monitoring on a machine using a user-level systemd timer.
| iamwsll/gpu-sentinel-feishu | 0xhassaan/nn-from-scratch | 3ks/embedoc | |
|---|---|---|---|
| Stars | 0 | 0 | — |
| Language | Python | Python | Python |
| Last pushed | — | — | 2023-06-08 |
| Maintenance | — | — | Dormant |
| Setup difficulty | moderate | moderate | hard |
| Complexity | 3/5 | 4/5 | 1/5 |
| Audience | ops devops | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires passwordless SSH access to the GPU server and a Feishu group webhook URL.
GPU Sentinel Feishu is a small Linux background service that watches NVIDIA GPU servers remotely and sends you a notification in Feishu, a team messaging app, whenever the workload on those GPUs changes. Instead of flooding you with constant status updates, it only alerts you when something actually happens: a new process starts using a GPU, an existing process finishes, or the very first snapshot of activity is recorded. It connects to your GPU server over SSH, so it does not need to be installed on the GPU machine itself, it sits on any Linux computer that can reach that server. It uses the standard nvidia-smi tool on the remote machine to collect information about GPU utilization, memory usage, temperature, power draw, and which processes are running. All of that history is stored in a local SQLite database, a single file with no separate database server required. The notifications it sends are formatted as Feishu interactive cards, which can include collapsible sections showing the full command each GPU process is running, plus system stats like CPU load and uptime. You set up a Feishu group bot, give its webhook URL to the installer, and the service takes care of the rest. Installation runs through a shell script that asks for your SSH connection details, the webhook URL, and a few preferences, then registers a user-level systemd timer, a Linux scheduler, so the check runs automatically without requiring administrator access. The project is written in Python and licensed under MIT.
A Linux service that watches remote NVIDIA GPU servers over SSH and sends Feishu notifications only when GPU workload actually changes.
Mainly Python. The stack also includes Python, SSH, systemd.
Free to use, modify, and distribute, even commercially, as long as you keep the original copyright and license notice.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly ops devops.
This repo across BitVibe Labs
Verify against the repo before relying on details.