 Chris Farhood
|
a226f0191c
|
feat: add Metrics page, remove app bar badge, fix sidebar label
- Add src/api/metrics.ts: Prometheus text parser + fetchGpuPluginMetrics()
fetching from Intel GPU device plugin pods (port 9090). Extracts engine
utilization (active/total ticks → %), boost frequency (MHz), VRAM and
system memory usage, cumulative energy (µJ).
- Add src/components/MetricsPage.tsx: per-card metrics display with inline
utilization bars, graceful fallback when enableMonitoring is not set.
- Register Metrics sidebar entry (mdi:chart-line) and route /intel-gpu/metrics.
- Remove registerAppBarAction and AppBarGpuBadge (colored info bubble).
- Fix sidebar parent label: 'Intel GPU' → 'intel-gpu'.
- Bump to v0.2.0; update artifacthub-pkg.yml with new archive URL and checksum.
Generated with [Claude Code](https://claude.ai/code)
via [Happy](https://happy.engineering)
Co-Authored-By: Claude <noreply@anthropic.com>
Co-Authored-By: Happy <yesreply@happy.engineering>
v0.2.0
|
2026-02-18 21:23:36 -05:00 |
|
|
Chris Farhood
|
41bf2aead4
|
feat: initial release of headlamp-intel-gpu-plugin v0.1.0
Adds a Headlamp plugin for Intel GPU device plugin visibility:
- Dedicated sidebar section: Overview, Device Plugins, GPU Nodes, GPU Pods
- Native Node detail page injection: GPU capacity, allocatable, utilization, active pods
- Native Pod detail page injection: per-container GPU resource requests/limits
- Native Nodes table: GPU Type and GPU Devices columns
- App bar health badge (hidden when plugin not installed)
- GpuDevicePlugin CRD monitoring (deviceplugin.intel.com/v1) with graceful
degradation when CRD is not present
- Supports discrete (i915), Xe, and integrated GPU nodes via node labels
- 48 unit tests, TypeScript clean, 28 kB production bundle
Generated with [Claude Code](https://claude.ai/code)
via [Happy](https://happy.engineering)
Co-Authored-By: Claude <noreply@anthropic.com>
Co-Authored-By: Happy <yesreply@happy.engineering>
|
2026-02-18 17:58:49 -05:00 |
|