fluidstack-gpu-dashboard/dashboard.yaml at master · bojiang3/fluidstack-gpu-dashboard · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
# dashboard.yaml — Kubernetes manifests for Fluidstack GPU Dashboard
# Apply with: kubectl apply -f dashboard.yaml  OR  python deploy.py

---
# ─── Deployment: GPU Dashboard (backend + frontend) ─────────────────────────
apiVersion: apps/v1
kind: Deployment
metadata:
  name: fluidstack-gpu-dashboard
  labels:
    app: fluidstack-gpu-dashboard
    hackathon: power-to-prefill
spec:
  replicas: 1
  selector:
    matchLabels:
      app: fluidstack-gpu-dashboard
  template:
    metadata:
      labels:
        app: fluidstack-gpu-dashboard
      annotations:
        prometheus.io/scrape: "true"
        prometheus.io/port: "8000"
        prometheus.io/path: "/prometheus"
    spec:
      containers:
        # ── FastAPI Backend ──
        - name: backend
          image: python:3.11-slim
          command:
            - sh
            - -c
            - |
              pip install fastapi uvicorn requests prometheus_client &&
              cd /app && python -m uvicorn app:app --host 0.0.0.0 --port 8000
          ports:
            - containerPort: 8000
              name: api
          env:
            - name: FLUIDSTACK_API_KEY
              valueFrom:
                secretKeyRef:
                  name: fluidstack-credentials
                  key: api-key
                  optional: true
            - name: NUM_MOCK_GPUS
              value: "4"
          resources:
            requests:
              cpu: "250m"
              memory: "256Mi"
            limits:
              cpu: "500m"
              memory: "512Mi"
              nvidia.com/gpu: "0"   # Set to 1+ when running on GPU nodes
          volumeMounts:
            - name: app-code
              mountPath: /app
          livenessProbe:
            httpGet:
              path: /
              port: 8000
            initialDelaySeconds: 30
            periodSeconds: 10
          readinessProbe:
            httpGet:
              path: /
              port: 8000
            initialDelaySeconds: 10
            periodSeconds: 5

        # ── Streamlit Frontend ──
        - name: frontend
          image: python:3.11-slim
          command:
            - sh
            - -c
            - |
              pip install streamlit plotly requests pandas numpy &&
              cd /app && streamlit run dashboard.py --server.port=8501 --server.headless=true --server.address=0.0.0.0
          ports:
            - containerPort: 8501
              name: ui
          resources:
            requests:
              cpu: "250m"
              memory: "256Mi"
            limits:
              cpu: "500m"
              memory: "512Mi"
          volumeMounts:
            - name: app-code
              mountPath: /app

      volumes:
        - name: app-code
          configMap:
            name: fluidstack-dashboard-code
            optional: true
        # NOTE: In production, mount app code via a proper container image
        # or use an initContainer to clone from git.

---
# ─── Service: NodePort for external access ──────────────────────────────────
apiVersion: v1
kind: Service
metadata:
  name: fluidstack-gpu-dashboard-svc
  labels:
    app: fluidstack-gpu-dashboard
spec:
  type: NodePort
  selector:
    app: fluidstack-gpu-dashboard
  ports:
    - name: api
      port: 8000
      targetPort: 8000
      nodePort: 30800
    - name: ui
      port: 8501
      targetPort: 8501
      nodePort: 30851

---
# ─── NVIDIA Device Plugin DaemonSet ─────────────────────────────────────────
# Required for GPU scheduling on Fluidstack/bare-metal/cloud k8s nodes.
# On Fluidstack managed clusters, this is typically pre-installed.
apiVersion: apps/v1
kind: DaemonSet
metadata:
  name: nvidia-device-plugin
  namespace: kube-system
  labels:
    app: nvidia-device-plugin
spec:
  selector:
    matchLabels:
      app: nvidia-device-plugin
  updateStrategy:
    type: RollingUpdate
  template:
    metadata:
      labels:
        app: nvidia-device-plugin
    spec:
      tolerations:
        - key: nvidia.com/gpu
          operator: Exists
          effect: NoSchedule
      priorityClassName: system-node-critical
      containers:
        - name: nvidia-device-plugin
          image: nvcr.io/nvidia/k8s-device-plugin:v0.15.0
          securityContext:
            allowPrivilegeEscalation: false
            capabilities:
              drop: ["ALL"]
          volumeMounts:
            - name: device-plugin
              mountPath: /var/lib/kubelet/device-plugins
      volumes:
        - name: device-plugin
          hostPath:
            path: /var/lib/kubelet/device-plugins