diff --git a/blog/hami-webui-v1-1-0-release/index.md b/blog/hami-webui-v1-1-0-release/index.md index be4b78cf..4d1e9d0b 100644 --- a/blog/hami-webui-v1-1-0-release/index.md +++ b/blog/hami-webui-v1-1-0-release/index.md @@ -10,7 +10,7 @@ image: /img/docs/en/userguide/webui-overview.png Managing GPU resources in Kubernetes has long been a "blind spot" for operators. You know GPUs are being used, but answering questions like "which node has idle capacity?", "is this workload actually utilizing its allocated GPU?", or "what's the overall cluster utilization trend?" often requires piecing together `kubectl get`, Prometheus PromQL, and log output. -Today, the HAMi community is introducing **[HAMi WebUI](https://github.com/Project-HAMi/HAMi-WebUI)** — an open-source GPU monitoring dashboard that puts your entire GPU cluster into a single, visual interface. +Today, the HAMi community is introducing **[HAMi WebUI](https://github.com/Project-HAMi/HAMi-WebUI)** - an open-source GPU monitoring dashboard that puts your entire GPU cluster into a single, visual interface. HAMi WebUI v1.1.0 is now available as the first official major release. @@ -20,7 +20,7 @@ Together with the core HAMi scheduler, WebUI completes the full loop: **from GPU ## The Challenge of GPU Monitoring in Kubernetes -[HAMi](https://github.com/Project-HAMi/HAMi), a [CNCF Sandbox project](https://www.cncf.io/projects/hami/), has long been focused on the scheduling and management layer of GPU resources in Kubernetes. The scheduler decides which GPU a workload gets — but once workloads are running, understanding what's happening at the resource level has been difficult. +[HAMi](https://github.com/Project-HAMi/HAMi), a [CNCF Sandbox project](https://www.cncf.io/projects/hami/), has long been focused on the scheduling and management layer of GPU resources in Kubernetes. The scheduler decides which GPU a workload gets - but once workloads are running, understanding what's happening at the resource level has been difficult. Consider a typical day-to-day scenario: @@ -28,7 +28,7 @@ Consider a typical day-to-day scenario: - An SRE receives an alert about high GPU memory usage and needs to pinpoint the source. - A cluster admin wants to compare GPU utilization across nodes to rebalance workloads. -Without a visual tool, each of these tasks requires switching between multiple terminal windows and manually correlating data from different sources. HAMi WebUI was built to solve exactly this problem — it is a purpose-built **GPU monitoring dashboard for Kubernetes**. +Without a visual tool, each of these tasks requires switching between multiple terminal windows and manually correlating data from different sources. HAMi WebUI was built to solve exactly this problem - it is a purpose-built **GPU monitoring dashboard for Kubernetes**. ## What Is HAMi WebUI? @@ -62,7 +62,7 @@ The cluster overview page provides a quick summary of the overall running status ![HAMi WebUI cluster overview dashboard showing GPU resource usage](/img/docs/en/userguide/webui-overview.png) -Key metrics support drill-down navigation — for example, clicking on "Schedulable" takes you directly to the node management page filtered by that state. You can grasp the global GPU cluster status on a single page. +Key metrics support drill-down navigation - for example, clicking on "Schedulable" takes you directly to the node management page filtered by that state. You can grasp the global GPU cluster status on a single page. ### Node Management @@ -74,7 +74,7 @@ On the node details page, you can further inspect GPU usage and task distributio ### GPU/Accelerator Management -From the accelerator perspective, WebUI supports fine-grained inspection of each GPU card — including allocation status, utilization, memory usage, and node association. +From the accelerator perspective, WebUI supports fine-grained inspection of each GPU card - including allocation status, utilization, memory usage, and node association. ![HAMi WebUI accelerator list showing per-GPU allocation, utilization, and memory usage](/img/docs/en/userguide/webui-accelerator-list.png) @@ -125,14 +125,14 @@ For detailed instructions, refer to the [HAMi WebUI Installation Guide](/docs/in The HAMi community has prepared comprehensive documentation: -- **[WebUI User Guide](/docs/userguide/hami-webui-user-guide)** — Learn how to use the cluster overview, node management, GPU management, and workload tracking features. -- **[WebUI Developer Guide](/docs/developers/hami-webui-development-guide)** — Understand the architecture, repository structure, local development setup, and coding conventions for contributing to WebUI. +- **[WebUI User Guide](/docs/userguide/hami-webui-user-guide)** - Learn how to use the cluster overview, node management, GPU management, and workload tracking features. +- **[WebUI Developer Guide](/docs/developers/hami-webui-development-guide)** - Understand the architecture, repository structure, local development setup, and coding conventions for contributing to WebUI. ## Join the HAMi Community HAMi WebUI is an active and evolving project. v1.1.0 already brings internationalization support (English and Chinese), multi-architecture container images, and broader device vendor compatibility. -We welcome contributions of all kinds — bug fixes, feature requests, documentation improvements, and new ideas. Check out the [HAMi-WebUI repository](https://github.com/Project-HAMi/HAMi-WebUI) and the [Developer Guide](/docs/developers/hami-webui-development-guide) to get started. +We welcome contributions of all kinds - bug fixes, feature requests, documentation improvements, and new ideas. Check out the [HAMi-WebUI repository](https://github.com/Project-HAMi/HAMi-WebUI) and the [Developer Guide](/docs/developers/hami-webui-development-guide) to get started. See the [v1.1.0 release notes](https://github.com/Project-HAMi/HAMi-WebUI/releases/tag/v1.1.0) for the full list of changes.