Merge dev into master

Author: gaurav00700

.gitignore

@@ -5,4 +5,6 @@
 *.txt
 *.rdb
 *.DS_Store
-*cache
+*cache
+*argo*
+*helm*

k8s/backend.yaml

@@ -0,0 +1,60 @@
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: backend
+  namespace: llm-backend
+  # annotations:
+  #   argocd-image-updater.argoproj.io/image-list: backend=ghcr.io/my-org/backend:latest
+  #   argocd-image-updater.argoproj.io/write-back-method: git
+spec:
+  replicas: 2
+  selector:
+    matchLabels:
+      app: backend
+  template:
+    metadata:
+      labels:
+        app: backend
+    spec:
+      containers:
+      - name: backend
+        image: gaurav00700/llm_app-backend:latest
+        # imagePullPolicy: Always  # Always pulls the latest image when the pod restarts
+        ports:
+        - containerPort: 8000
+        env:
+        - name: LLM_NAME
+          valueFrom:
+            configMapKeyRef:
+              name: llm-config
+              key: LLM_NAME
+        - name: LLM_HOST_NAME
+          value: "llm.llm-model.svc.cluster.local"  # <service-name>.<namespace>.svc.cluster.local
+        - name: LLM_PORT
+          value: "11434"
+        - name: REDIS_HOST_NAME
+          value: "redis.llm-redis.svc.cluster.local" # <service-name>.<namespace>.svc.cluster.local
+        - name: REDIS_PORT
+          value: "6379"
+        volumeMounts:
+        - mountPath: /workdir/data
+          name: data-volume
+      volumes:  # Call the volume mount
+      - name: data-volume
+        persistentVolumeClaim:
+          claimName: data-pvc
+
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: backend
+  namespace: llm-backend
+spec:
+  type: ClusterIP
+  selector:
+    app: backend
+  ports:
+  - protocol: TCP
+    port: 8000
+    targetPort: 8000

k8s/configmap.yaml

@@ -0,0 +1,16 @@
+apiVersion: v1
+kind: ConfigMap
+metadata:
+  name: llm-config
+  namespace: llm-backend
+data:
+  LLM_NAME: "qwen2.5:0.5b"
+
+---
+apiVersion: v1
+kind: ConfigMap
+metadata:
+  name: llm-config
+  namespace: llm-model
+data:
+  LLM_NAME: "qwen2.5:0.5b"

k8s/frontend.yaml

@@ -0,0 +1,49 @@
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: frontend    # service name
+  namespace: llm-frontend   # namespace name
+  # annotations:
+  #   argocd-image-updater.argoproj.io/image-list: frontend=ghcr.io/my-org/frontend:latest
+  #   argocd-image-updater.argoproj.io/write-back-method: git
+spec:
+  replicas: 2
+  selector:
+    matchLabels:
+      app: frontend
+  template:
+    metadata:
+      labels:
+        app: frontend
+    spec:
+      containers:
+      - name: frontend
+        image: gaurav00700/llm_app-frontend:latest
+        # imagePullPolicy: Always  # Always pulls the latest image when the pod restarts
+        ports:
+        - containerPort: 8500
+        env:
+        - name: BACKEND_ENDPNT
+          value: "http://backend.llm-backend.svc.cluster.local:8000/ask"
+        volumeMounts:
+        - mountPath: /workdir/data
+          name: data-volume
+      volumes:  # Call the volume mount
+      - name: data-volume
+        persistentVolumeClaim:
+          claimName: data-pvc   # Matches PVC name in the same namespace
+
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: frontend
+  namespace: llm-frontend
+spec:
+  type: LoadBalancer
+  selector:
+    app: frontend
+  ports:
+  - protocol: TCP
+    port: 8500  # host port
+    targetPort: 8500  # pod port

k8s/model.yaml

@@ -0,0 +1,55 @@
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: llm
+  namespace: llm-model
+  # annotations:
+  #   argocd-image-updater.argoproj.io/image-list: model=ghcr.io/my-org/model:latest
+  #   argocd-image-updater.argoproj.io/write-back-method: git
+spec:
+  replicas: 1
+  selector:
+    matchLabels:
+      app: llm
+  template:
+    metadata:
+      labels:
+        app: llm
+    spec:
+      containers:
+      - name: llm
+        image: gaurav00700/llm_app-llm:latest
+        # imagePullPolicy: Always  # Always pulls the latest image when the pod restarts
+        ports:
+        - containerPort: 11434
+        env:
+        - name: LLM_NAME
+          valueFrom:
+            configMapKeyRef:
+              name: llm-config
+              key: LLM_NAME
+        volumeMounts:
+        - mountPath: /root/.ollama
+          name: llm-volume
+        resources:
+          limits:
+            nvidia.com/gpu: 1  # GPU qty
+      volumes:  # Call the volume mount
+      - name: llm-volume
+        persistentVolumeClaim:
+          claimName: llm-pvc
+
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: llm
+  namespace: llm-model
+spec:
+  type: ClusterIP
+  selector:
+    app: llm
+  ports:
+  - protocol: TCP
+    port: 11434
+    targetPort: 11434

k8s/namespaces.yaml

@@ -0,0 +1,22 @@
+apiVersion: v1
+kind: Namespace
+metadata:
+  name: llm-frontend
+
+---
+apiVersion: v1
+kind: Namespace
+metadata:
+  name: llm-backend
+
+---
+apiVersion: v1
+kind: Namespace
+metadata:
+  name: llm-redis
+  
+---
+apiVersion: v1
+kind: Namespace
+metadata:
+  name: llm-model

k8s/redis.yaml

@@ -0,0 +1,43 @@
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: redis
+  namespace: llm-redis
+spec:
+  replicas: 1
+  selector:
+    matchLabels:
+      app: redis
+  template:
+    metadata:
+      labels:
+        app: redis
+    spec:
+      containers:
+      - name: redis
+        image: redis:latest
+        # imagePullPolicy: Always  # Always pulls the latest image when the pod restarts
+        ports:
+        - containerPort: 6379
+        volumeMounts:
+        - mountPath: /data
+          name: redis-volume
+      volumes:  # Call the volume mount
+      - name: redis-volume
+        persistentVolumeClaim:
+          claimName: redis-pvc
+
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: redis
+  namespace: llm-redis
+spec:
+  type: ClusterIP
+  selector:
+    app: redis
+  ports:
+  - protocol: TCP
+    port: 6379
+    targetPort: 6379

k8s/volumes.yaml

@@ -0,0 +1,68 @@
+apiVersion: v1
+kind: PersistentVolume
+metadata:
+  name: shared-pv
+spec:
+  capacity:
+    storage: 5Gi
+  accessModes:
+    - ReadWriteOnce  # Allow multiple pods to read/write [ReadWriteMany, ReadWriteOnce]
+  persistentVolumeReclaimPolicy: Retain # Data remains after PVC deletion
+  hostPath:
+    path: "/data"  # Must match the mounted path inside Minikube
+
+---
+apiVersion: v1
+kind: PersistentVolumeClaim
+metadata:
+  name: data-pvc
+  namespace: llm-frontend
+spec:
+  accessModes:
+    - ReadWriteOnce  # Allow multiple pods to read/write [ReadWriteMany, ReadWriteOnce]
+  resources:
+    requests:
+      storage: 100Mi  # Gi, Mi
+  # volumeName: shared-pv  # binds to PV
+  # storageClassName: ""  # Manually created PV, so no storage class
+
+---
+apiVersion: v1
+kind: PersistentVolumeClaim
+metadata:
+  name: data-pvc
+  namespace: llm-backend
+spec:
+  accessModes:
+    - ReadWriteOnce
+  resources:
+    requests:
+      storage: 100Mi
+  # volumeName: shared-pv  # binds to PV
+  # storageClassName: ""  # Must be empty to use a manually created PV
+
+---
+apiVersion: v1
+kind: PersistentVolumeClaim
+metadata:
+  name: redis-pvc
+  namespace: llm-redis
+spec:
+  accessModes:
+    - ReadWriteMany
+  resources:
+    requests:
+      storage: 100Mi
+
+---
+apiVersion: v1
+kind: PersistentVolumeClaim
+metadata:
+  name: llm-pvc
+  namespace: llm-model
+spec:
+  accessModes:
+    - ReadWriteMany
+  resources:
+    requests:
+      storage: 5Gi