Some checks failed
Tests / Go Tests (push) Has been cancelled
Tests / Python Tests (push) Has been cancelled
Tests / Integration Tests (push) Has been cancelled
Tests / Go Lint (push) Has been cancelled
Tests / Python Lint (push) Has been cancelled
Tests / Security Scan (push) Has been cancelled
Tests / All Checks Passed (push) Has been cancelled
Security Scanning / Secret Scanning (push) Has been cancelled
Security Scanning / Dependency Vulnerability Scan (push) Has been cancelled
Security Scanning / Go Security Scan (push) Has been cancelled
Security Scanning / Python Security Scan (push) Has been cancelled
Security Scanning / Node.js Security Scan (push) Has been cancelled
Security Scanning / Docker Image Security (push) Has been cancelled
Security Scanning / Security Summary (push) Has been cancelled
CI/CD Pipeline / Go Tests (push) Has been cancelled
CI/CD Pipeline / Python Tests (push) Has been cancelled
CI/CD Pipeline / Website Tests (push) Has been cancelled
CI/CD Pipeline / Linting (push) Has been cancelled
CI/CD Pipeline / Security Scan (push) Has been cancelled
CI/CD Pipeline / Docker Build & Push (push) Has been cancelled
CI/CD Pipeline / Integration Tests (push) Has been cancelled
CI/CD Pipeline / Deploy to Staging (push) Has been cancelled
CI/CD Pipeline / Deploy to Production (push) Has been cancelled
CI/CD Pipeline / CI Summary (push) Has been cancelled
ci/woodpecker/manual/build-ci-image Pipeline was successful
ci/woodpecker/manual/main Pipeline failed
All services: admin-v2, studio-v2, website, ai-compliance-sdk, consent-service, klausur-service, voice-service, and infrastructure. Large PDFs and compiled binaries excluded via .gitignore.
366 lines
11 KiB
TypeScript
366 lines
11 KiB
TypeScript
'use client'
|
|
|
|
import { useState } from 'react'
|
|
import AdminLayout from '@/components/admin/AdminLayout'
|
|
import {
|
|
WizardStepper,
|
|
WizardNavigation,
|
|
EducationCard,
|
|
ArchitectureContext,
|
|
TestRunner,
|
|
TestSummary,
|
|
type WizardStep,
|
|
type TestCategoryResult,
|
|
type FullTestResults,
|
|
type EducationContent,
|
|
type ArchitectureContextType,
|
|
} from '@/components/wizard'
|
|
|
|
// ==============================================
|
|
// Constants
|
|
// ==============================================
|
|
|
|
const BACKEND_URL = process.env.NEXT_PUBLIC_BACKEND_URL || 'http://localhost:8000'
|
|
|
|
const STEPS: WizardStep[] = [
|
|
{ id: 'welcome', name: 'Willkommen', icon: '👋', status: 'pending' },
|
|
{ id: 'api-health', name: 'API Status', icon: '💚', status: 'pending', category: 'api-health' },
|
|
{ id: 'detection', name: 'GPU Erkennung', icon: '🔍', status: 'pending', category: 'detection' },
|
|
{ id: 'cloud', name: 'Cloud GPU', icon: '☁️', status: 'pending', category: 'cloud' },
|
|
{ id: 'summary', name: 'Zusammenfassung', icon: '📊', status: 'pending' },
|
|
]
|
|
|
|
const EDUCATION_CONTENT: Record<string, EducationContent> = {
|
|
'welcome': {
|
|
title: 'Willkommen zum GPU Wizard',
|
|
content: [
|
|
'GPUs beschleunigen KI-Workloads um das 10-100fache.',
|
|
'',
|
|
'BreakPilot unterstuetzt:',
|
|
'• NVIDIA GPUs (CUDA) - GeForce, Tesla, A100',
|
|
'• AMD GPUs (ROCm) - Radeon, Instinct',
|
|
'• Cloud GPUs (vast.ai) - On-Demand Miete',
|
|
'',
|
|
'Anwendungsfaelle:',
|
|
'• LLM Inference (lokale Modelle)',
|
|
'• Embedding-Generierung',
|
|
'• Fine-Tuning von Modellen',
|
|
'• OCR und Bildverarbeitung',
|
|
],
|
|
},
|
|
'api-health': {
|
|
title: 'GPU Admin API - Verwaltungsschnittstelle',
|
|
content: [
|
|
'Die GPU API verwaltet lokale und Cloud-GPUs.',
|
|
'',
|
|
'Endpunkte:',
|
|
'• /api/gpu/status - Aktueller GPU-Status',
|
|
'• /api/gpu/jobs - Laufende GPU-Jobs',
|
|
'• /api/gpu/metrics - Auslastung und Temperatur',
|
|
'',
|
|
'vast.ai Integration:',
|
|
'• /vast/instances - Gemietete Instanzen',
|
|
'• /vast/offers - Verfuegbare Angebote',
|
|
'• /vast/start - Instanz starten',
|
|
],
|
|
},
|
|
'detection': {
|
|
title: 'GPU Hardware-Erkennung',
|
|
content: [
|
|
'Automatische Erkennung verfuegbarer GPUs.',
|
|
'',
|
|
'NVIDIA (nvidia-smi):',
|
|
'• GPU-Modell und VRAM',
|
|
'• Treiber-Version',
|
|
'• Aktuelle Auslastung',
|
|
'• Temperatur und Stromverbrauch',
|
|
'',
|
|
'AMD (rocm-smi):',
|
|
'• ROCm-Version',
|
|
'• GPU-Modell',
|
|
'• Memory-Statistiken',
|
|
'',
|
|
'PyTorch CUDA Check:',
|
|
'• torch.cuda.is_available()',
|
|
'• torch.cuda.device_count()',
|
|
],
|
|
},
|
|
'cloud': {
|
|
title: 'Cloud GPUs mit vast.ai',
|
|
content: [
|
|
'vast.ai bietet guenstige Cloud-GPUs auf Abruf.',
|
|
'',
|
|
'Vorteile:',
|
|
'• 3-10x guenstiger als AWS/GCP',
|
|
'• Peer-to-Peer Marktplatz',
|
|
'• Minutengenaue Abrechnung',
|
|
'• Zugriff auf A100, H100, etc.',
|
|
'',
|
|
'Typische Preise:',
|
|
'• RTX 3090: ~$0.20/Stunde',
|
|
'• A100 40GB: ~$1.00/Stunde',
|
|
'• H100: ~$2.50/Stunde',
|
|
'',
|
|
'Konfiguration: VAST_API_KEY Umgebungsvariable',
|
|
],
|
|
},
|
|
'summary': {
|
|
title: 'Test-Zusammenfassung',
|
|
content: [
|
|
'Hier sehen Sie eine Uebersicht aller durchgefuehrten Tests:',
|
|
'• Lokale GPU-Erkennung',
|
|
'• Cloud-GPU Verfuegbarkeit',
|
|
'• API-Status',
|
|
],
|
|
},
|
|
}
|
|
|
|
const ARCHITECTURE_CONTEXTS: Record<string, ArchitectureContextType> = {
|
|
'api-health': {
|
|
layer: 'api',
|
|
services: ['backend'],
|
|
dependencies: ['PostgreSQL', 'vast.ai API'],
|
|
dataFlow: ['Browser', 'FastAPI', 'GPU Manager', 'PostgreSQL'],
|
|
},
|
|
'detection': {
|
|
layer: 'service',
|
|
services: ['backend'],
|
|
dependencies: ['nvidia-smi', 'rocm-smi', 'PyTorch'],
|
|
dataFlow: ['FastAPI', 'Subprocess', 'GPU Driver', 'Hardware'],
|
|
},
|
|
'cloud': {
|
|
layer: 'service',
|
|
services: ['backend'],
|
|
dependencies: ['vast.ai API', 'SSH', 'Docker'],
|
|
dataFlow: ['FastAPI', 'vast.ai API', 'Cloud Instance', 'GPU'],
|
|
},
|
|
}
|
|
|
|
// ==============================================
|
|
// Main Component
|
|
// ==============================================
|
|
|
|
export default function GPUWizardPage() {
|
|
const [currentStep, setCurrentStep] = useState(0)
|
|
const [steps, setSteps] = useState<WizardStep[]>(STEPS)
|
|
const [categoryResults, setCategoryResults] = useState<Record<string, TestCategoryResult>>({})
|
|
const [fullResults, setFullResults] = useState<FullTestResults | null>(null)
|
|
const [isLoading, setIsLoading] = useState(false)
|
|
const [error, setError] = useState<string | null>(null)
|
|
|
|
const currentStepData = steps[currentStep]
|
|
const isTestStep = currentStepData?.category !== undefined
|
|
const isWelcome = currentStepData?.id === 'welcome'
|
|
const isSummary = currentStepData?.id === 'summary'
|
|
|
|
const runCategoryTest = async (category: string) => {
|
|
setIsLoading(true)
|
|
setError(null)
|
|
|
|
try {
|
|
const response = await fetch(`${BACKEND_URL}/api/admin/gpu-tests/${category}`, {
|
|
method: 'POST',
|
|
})
|
|
|
|
if (!response.ok) {
|
|
throw new Error(`HTTP ${response.status}: ${response.statusText}`)
|
|
}
|
|
|
|
const result: TestCategoryResult = await response.json()
|
|
setCategoryResults((prev) => ({ ...prev, [category]: result }))
|
|
|
|
setSteps((prev) =>
|
|
prev.map((step) =>
|
|
step.category === category
|
|
? { ...step, status: result.failed === 0 ? 'completed' : 'failed' }
|
|
: step
|
|
)
|
|
)
|
|
} catch (err) {
|
|
setError(err instanceof Error ? err.message : 'Unbekannter Fehler')
|
|
} finally {
|
|
setIsLoading(false)
|
|
}
|
|
}
|
|
|
|
const runAllTests = async () => {
|
|
setIsLoading(true)
|
|
setError(null)
|
|
|
|
try {
|
|
const response = await fetch(`${BACKEND_URL}/api/admin/gpu-tests/run-all`, {
|
|
method: 'POST',
|
|
})
|
|
|
|
if (!response.ok) {
|
|
throw new Error(`HTTP ${response.status}: ${response.statusText}`)
|
|
}
|
|
|
|
const results: FullTestResults = await response.json()
|
|
setFullResults(results)
|
|
|
|
setSteps((prev) =>
|
|
prev.map((step) => {
|
|
if (step.category) {
|
|
const catResult = results.categories.find((c) => c.category === step.category)
|
|
if (catResult) {
|
|
return { ...step, status: catResult.failed === 0 ? 'completed' : 'failed' }
|
|
}
|
|
}
|
|
return step
|
|
})
|
|
)
|
|
|
|
const newCategoryResults: Record<string, TestCategoryResult> = {}
|
|
results.categories.forEach((cat) => {
|
|
newCategoryResults[cat.category] = cat
|
|
})
|
|
setCategoryResults(newCategoryResults)
|
|
} catch (err) {
|
|
setError(err instanceof Error ? err.message : 'Unbekannter Fehler')
|
|
} finally {
|
|
setIsLoading(false)
|
|
}
|
|
}
|
|
|
|
const goToNext = () => {
|
|
if (currentStep < steps.length - 1) {
|
|
setSteps((prev) =>
|
|
prev.map((step, idx) =>
|
|
idx === currentStep && step.status === 'pending'
|
|
? { ...step, status: 'completed' }
|
|
: step
|
|
)
|
|
)
|
|
setCurrentStep((prev) => prev + 1)
|
|
}
|
|
}
|
|
|
|
const goToPrev = () => {
|
|
if (currentStep > 0) {
|
|
setCurrentStep((prev) => prev - 1)
|
|
}
|
|
}
|
|
|
|
const handleStepClick = (index: number) => {
|
|
if (index <= currentStep || steps[index - 1]?.status !== 'pending') {
|
|
setCurrentStep(index)
|
|
}
|
|
}
|
|
|
|
return (
|
|
<AdminLayout
|
|
title="GPU Wizard"
|
|
description="Interaktives Lernen und Testen der GPU-Infrastruktur"
|
|
>
|
|
{/* Header */}
|
|
<div className="bg-white rounded-lg shadow p-4 mb-6 flex items-center justify-between">
|
|
<div className="flex items-center">
|
|
<span className="text-3xl mr-3">🎮</span>
|
|
<div>
|
|
<h2 className="text-lg font-bold text-gray-800">GPU Infrastructure Wizard</h2>
|
|
<p className="text-sm text-gray-600">CUDA, ROCm & vast.ai</p>
|
|
</div>
|
|
</div>
|
|
<a href="/admin/gpu" className="text-blue-600 hover:text-blue-800 text-sm">
|
|
← Zurueck zu GPU Management
|
|
</a>
|
|
</div>
|
|
|
|
{/* Stepper */}
|
|
<div className="bg-white rounded-lg shadow p-6 mb-6">
|
|
<WizardStepper steps={steps} currentStep={currentStep} onStepClick={handleStepClick} />
|
|
</div>
|
|
|
|
{/* Content */}
|
|
<div className="bg-white rounded-lg shadow p-6">
|
|
<div className="flex items-center mb-6">
|
|
<span className="text-3xl mr-3">{currentStepData?.icon}</span>
|
|
<div>
|
|
<h2 className="text-xl font-bold text-gray-800">
|
|
Schritt {currentStep + 1}: {currentStepData?.name}
|
|
</h2>
|
|
<p className="text-gray-500 text-sm">
|
|
{currentStep + 1} von {steps.length}
|
|
</p>
|
|
</div>
|
|
</div>
|
|
|
|
<EducationCard content={EDUCATION_CONTENT[currentStepData?.id || '']} />
|
|
|
|
{isTestStep && currentStepData?.category && ARCHITECTURE_CONTEXTS[currentStepData.category] && (
|
|
<ArchitectureContext
|
|
context={ARCHITECTURE_CONTEXTS[currentStepData.category]}
|
|
currentStep={currentStepData.name}
|
|
/>
|
|
)}
|
|
|
|
{error && (
|
|
<div className="bg-red-50 border border-red-200 text-red-700 rounded-lg p-4 mb-6">
|
|
<strong>Fehler:</strong> {error}
|
|
</div>
|
|
)}
|
|
|
|
{isWelcome && (
|
|
<div className="text-center py-8">
|
|
<button
|
|
onClick={goToNext}
|
|
className="bg-blue-600 text-white px-8 py-3 rounded-lg font-medium hover:bg-blue-700 transition-colors"
|
|
>
|
|
Wizard starten
|
|
</button>
|
|
</div>
|
|
)}
|
|
|
|
{isTestStep && currentStepData?.category && (
|
|
<TestRunner
|
|
category={currentStepData.category}
|
|
categoryResult={categoryResults[currentStepData.category]}
|
|
isLoading={isLoading}
|
|
onRunTests={() => runCategoryTest(currentStepData.category!)}
|
|
/>
|
|
)}
|
|
|
|
{isSummary && (
|
|
<div>
|
|
{!fullResults ? (
|
|
<div className="text-center py-8">
|
|
<p className="text-gray-600 mb-4">
|
|
Fuehren Sie alle Tests aus um eine Zusammenfassung zu sehen.
|
|
</p>
|
|
<button
|
|
onClick={runAllTests}
|
|
disabled={isLoading}
|
|
className={`px-6 py-3 rounded-lg font-medium transition-colors ${
|
|
isLoading
|
|
? 'bg-gray-400 cursor-not-allowed'
|
|
: 'bg-blue-600 text-white hover:bg-blue-700'
|
|
}`}
|
|
>
|
|
{isLoading ? 'Alle Tests laufen...' : 'Alle Tests ausfuehren'}
|
|
</button>
|
|
</div>
|
|
) : (
|
|
<TestSummary results={fullResults} />
|
|
)}
|
|
</div>
|
|
)}
|
|
|
|
<WizardNavigation
|
|
currentStep={currentStep}
|
|
totalSteps={steps.length}
|
|
onPrev={goToPrev}
|
|
onNext={goToNext}
|
|
showNext={!isSummary}
|
|
isLoading={isLoading}
|
|
/>
|
|
</div>
|
|
|
|
<div className="text-center text-gray-500 text-sm mt-6">
|
|
Diese Tests pruefen die GPU-Infrastruktur.
|
|
Bei Fragen wenden Sie sich an das DevOps-Team.
|
|
</div>
|
|
</AdminLayout>
|
|
)
|
|
}
|