feat(infra): prod conf

.github/workflows/deploy-prod.yml

@@ -0,0 +1,76 @@
+name: Deploy to k8s
+on:
+  # run this workflow manually from the Actions tab
+  workflow_dispatch:
+
+jobs:
+  build-and-publish:
+    runs-on: 
+      group: cpu-high
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+
+      - name: Login to Registry
+        uses: docker/login-action@v3
+        with:
+          username: ${{ secrets.DOCKER_INTERNAL_USERNAME }}
+          password: ${{ secrets.DOCKER_INTERNAL_PASSWORD }}
+
+      - name: Docker metadata
+        id: meta
+        uses: docker/metadata-action@v5
+        with:
+          images: |
+            registry.internal.huggingface.tech/deepsite/deepsite
+          tags: |
+            type=raw,value=latest,enable={{is_default_branch}}
+            type=sha,enable=true,prefix=sha-,format=short,sha-len=8
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+
+      - name: Inject slug/short variables
+        uses: rlespinasse/github-slug-action@v4.5.0
+
+      - name: Build and Publish image
+        uses: docker/build-push-action@v5
+        with:
+          context: .
+          file: Dockerfile
+          push: ${{ github.event_name != 'pull_request' }}
+          tags: ${{ steps.meta.outputs.tags }}
+          labels: ${{ steps.meta.outputs.labels }}
+          platforms: linux/amd64
+          cache-to: type=gha,mode=max,scope=amd64
+          cache-from: type=gha,scope=amd64
+          provenance: false
+
+  deploy:
+    name: Deploy on prod
+    runs-on: ubuntu-latest
+    needs: ["build-and-publish"]
+    steps:
+      - name: Inject slug/short variables
+        uses: rlespinasse/github-slug-action@v4.5.0
+
+      - name: Gen values
+        run: |
+          VALUES=$(cat <<-END
+          image:
+            tag: "sha-${{ env.GITHUB_SHA_SHORT }}"
+          END
+          )
+          echo "VALUES=$(echo "$VALUES" | yq -o=json | jq tostring)" >> $GITHUB_ENV
+
+      - name: Deploy on infra-deployments
+        uses: aurelien-baudet/workflow-dispatch@v2
+        with:
+          workflow: Update application single value
+          repo: huggingface/infra-deployments
+          wait-for-completion: true
+          wait-for-completion-interval: 10s
+          display-workflow-run-url-interval: 10s
+          ref: refs/heads/main
+          token: ${{ secrets.GIT_TOKEN_INFRA_DEPLOYMENT }}
+          inputs: '{"path": "hub/deepsite/deepsite.yaml", "value": ${{ env.VALUES }}, "url": "${{ github.event.head_commit.url }}"}'

app/api/healthcheck/route.ts

@@ -0,0 +1,5 @@
+import { NextResponse } from "next/server";
+
+export async function GET() {
+  return NextResponse.json({ status: "ok" }, { status: 200 });
+}

chart/Chart.yaml

@@ -0,0 +1,5 @@
+apiVersion: v2
+name: deepsite
+version: 0.0.0-latest
+type: application
+icon: https://huggingface.co/front/assets/huggingface_logo-noborder.svg

chart/env/prod.yaml

@@ -0,0 +1,53 @@
+nodeSelector:
+  role-deepsite: "true"
+
+tolerations:
+  - key: "huggingface.co/deepsite"
+    operator: "Equal"
+    value: "true"
+    effect: "NoSchedule"
+
+serviceAccount:
+  enabled: true
+  create: true
+  name: deepsite-prod
+
+ingress:
+  path: "/"
+  annotations:
+    alb.ingress.kubernetes.io/healthcheck-path: "/healthcheck"
+    alb.ingress.kubernetes.io/listen-ports: "[{\"HTTP\": 80}, {\"HTTPS\": 443}]"
+    alb.ingress.kubernetes.io/load-balancer-name: "hub-utils-prod-cloudfront"
+    alb.ingress.kubernetes.io/group.name: "hub-utils-prod-cloudfront"
+    alb.ingress.kubernetes.io/scheme: "internal"
+    alb.ingress.kubernetes.io/ssl-redirect: "443"
+    alb.ingress.kubernetes.io/tags: "Env=prod,Project=hub,Terraform=true"
+    alb.ingress.kubernetes.io/target-group-attributes: deregistration_delay.timeout_seconds=30
+    alb.ingress.kubernetes.io/target-type: "ip"
+    alb.ingress.kubernetes.io/certificate-arn: "arn:aws:acm:us-east-1:707930574880:certificate/5b25b145-75db-4837-b9f3-7f238ba8a9c7,arn:aws:acm:us-east-1:707930574880:certificate/bfdf509c-f44b-400f-b9e1-6f7a861abe91"
+    kubernetes.io/ingress.class: "alb"
+
+ingressInternal:
+  enabled: false
+
+envVars:
+  NEXTAUTH_URL=https://deepsite.hf.co/api/auth
+
+infisical:
+  enabled: true
+  env: "prod-us-east-1"
+
+autoscaling:
+  enabled: true
+  minReplicas: 1
+  maxReplicas: 10
+  targetMemoryUtilizationPercentage: "50"
+  targetCPUUtilizationPercentage: "50"
+
+resources:
+  requests:
+    cpu: 2
+    memory: 4Gi
+  limits:
+    cpu: 4
+    memory: 8Gi

chart/templates/_helpers.tpl

@@ -0,0 +1,22 @@
+{{- define "name" -}}
+{{- default $.Release.Name | trunc 63 | trimSuffix "-" -}}
+{{- end -}}
+
+{{- define "app.name" -}}
+chat-ui
+{{- end -}}
+
+{{- define "labels.standard" -}}
+release: {{ $.Release.Name | quote }}
+heritage: {{ $.Release.Service | quote }}
+chart: "{{ include "name" . }}"
+app: "{{ include "app.name" . }}"
+{{- end -}}
+
+{{- define "labels.resolver" -}}
+release: {{ $.Release.Name | quote }}
+heritage: {{ $.Release.Service | quote }}
+chart: "{{ include "name" . }}"
+app: "{{ include "app.name" . }}-resolver"
+{{- end -}}
+

chart/templates/config.yaml

@@ -0,0 +1,10 @@
+apiVersion: v1
+kind: ConfigMap
+metadata:
+  labels: {{ include "labels.standard" . | nindent 4 }}
+  name: {{ include "name" . }}
+  namespace: {{ .Release.Namespace }}
+data:
+  {{- range $key, $value := $.Values.envVars }}
+  {{ $key }}: {{ $value | quote }}
+  {{- end }}

chart/templates/deployment.yaml

@@ -0,0 +1,81 @@
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  labels: {{ include "labels.standard" . | nindent 4 }}
+  name: {{ include "name" . }}
+  namespace: {{ .Release.Namespace }}
+  {{- if .Values.infisical.enabled }}
+  annotations:
+    secrets.infisical.com/auto-reload: "true"
+  {{- end }}
+spec:
+  progressDeadlineSeconds: 600
+  {{- if not $.Values.autoscaling.enabled }}
+  replicas: {{ .Values.replicas }}
+  {{- end }}
+  revisionHistoryLimit: 10
+  selector:
+    matchLabels: {{ include "labels.standard" . | nindent 6 }}
+  strategy:
+    rollingUpdate:
+      maxSurge: 25%
+      maxUnavailable: 25%
+    type: RollingUpdate
+  template:
+    metadata:
+      labels: {{ include "labels.standard" . | nindent 8 }}
+      annotations:
+        checksum/config: {{ include (print $.Template.BasePath "/config.yaml") . | sha256sum }}
+        {{- if $.Values.envVars.NODE_LOG_STRUCTURED_DATA }}
+        co.elastic.logs/json.expand_keys: "true"
+        {{- end }}
+    spec:
+      {{- if .Values.serviceAccount.enabled }}
+      serviceAccountName: "{{ .Values.serviceAccount.name | default (include "name" .) }}"
+      {{- end }}
+      containers:
+        - name: chat-ui
+          image: "{{ .Values.image.repository }}/{{ .Values.image.name }}:{{ .Values.image.tag }}"
+          imagePullPolicy: {{ .Values.image.pullPolicy }}
+          readinessProbe:
+            failureThreshold: 30
+            periodSeconds: 10
+            httpGet:
+              path: {{ $.Values.envVars.APP_BASE | default "" }}/healthcheck
+              port: {{ $.Values.envVars.APP_PORT | default 3001 | int }}
+          livenessProbe:
+            failureThreshold: 30
+            periodSeconds: 10
+            httpGet:
+              path: {{ $.Values.envVars.APP_BASE | default "" }}/healthcheck
+              port: {{ $.Values.envVars.APP_PORT | default 3001 | int }}
+          ports:
+            - containerPort: {{ $.Values.envVars.APP_PORT | default 3001 | int }}
+              name: http
+              protocol: TCP
+            {{- if eq "true" $.Values.envVars.METRICS_ENABLED }}
+            - containerPort: {{ $.Values.envVars.METRICS_PORT | default 5565 | int }}
+              name: metrics
+              protocol: TCP
+            {{- end }}
+          resources: {{ toYaml .Values.resources | nindent 12 }}
+          {{- with $.Values.extraEnv }}
+          env:
+            {{- toYaml . | nindent 14 }}
+          {{- end }}
+          envFrom:
+            - configMapRef:
+                name: {{ include "name" . }}
+          {{- if $.Values.infisical.enabled }}
+            - secretRef:
+                name: {{ include "name" $ }}-secs
+          {{- end }}
+          {{- with $.Values.extraEnvFrom }}
+            {{- toYaml . | nindent 14 }}
+          {{- end }}
+      nodeSelector: {{ toYaml .Values.nodeSelector | nindent 8 }}
+      tolerations: {{ toYaml .Values.tolerations | nindent 8 }}
+      volumes:
+        - name: config
+          configMap:
+            name: {{ include "name" . }}

chart/templates/hpa.yaml

@@ -0,0 +1,45 @@
+{{- if $.Values.autoscaling.enabled }}
+apiVersion: autoscaling/v2
+kind: HorizontalPodAutoscaler
+metadata:
+  labels: {{ include "labels.standard" . | nindent 4 }}
+  name: {{ include "name" . }}
+  namespace: {{ .Release.Namespace }}
+spec:
+  scaleTargetRef:
+    apiVersion: apps/v1
+    kind: Deployment
+    name: {{ include "name" . }}
+  minReplicas: {{ $.Values.autoscaling.minReplicas }}
+  maxReplicas: {{ $.Values.autoscaling.maxReplicas }}
+  metrics:
+    {{- if ne "" $.Values.autoscaling.targetMemoryUtilizationPercentage }}
+    - type: Resource
+      resource:
+        name: memory
+        target:
+          type: Utilization
+          averageUtilization: {{ $.Values.autoscaling.targetMemoryUtilizationPercentage | int }}
+    {{- end }}
+    {{- if ne "" $.Values.autoscaling.targetCPUUtilizationPercentage }}
+    - type: Resource
+      resource:
+        name: cpu
+        target:
+          type: Utilization
+          averageUtilization: {{ $.Values.autoscaling.targetCPUUtilizationPercentage | int }}
+    {{- end }}
+  behavior:
+    scaleDown:
+      stabilizationWindowSeconds: 600
+      policies:
+        - type: Percent
+          value: 10
+          periodSeconds: 60
+    scaleUp:
+      stabilizationWindowSeconds: 0
+      policies:
+        - type: Pods
+          value: 1
+          periodSeconds: 30
+{{- end }}

chart/templates/infisical.yaml

@@ -0,0 +1,24 @@
+{{- if .Values.infisical.enabled }}
+apiVersion: secrets.infisical.com/v1alpha1
+kind: InfisicalSecret
+metadata:
+  name: {{ include "name" $ }}-infisical-secret
+  namespace: {{ $.Release.Namespace }}
+spec:
+  authentication:
+    universalAuth:
+      credentialsRef:
+        secretName: {{ .Values.infisical.operatorSecretName | quote }}
+        secretNamespace: {{ .Values.infisical.operatorSecretNamespace | quote }}
+      secretsScope:
+        envSlug: {{ .Values.infisical.env | quote }}
+        projectSlug: {{ .Values.infisical.project | quote }}
+        secretsPath: /
+  hostAPI: {{ .Values.infisical.url | quote }}
+  managedSecretReference:
+    creationPolicy: Owner
+    secretName: {{ include "name" $ }}-secs
+    secretNamespace: {{ .Release.Namespace | quote }}
+    secretType: Opaque
+  resyncInterval: {{ .Values.infisical.resyncInterval }}
+{{- end }}

chart/templates/ingress-internal.yaml

@@ -0,0 +1,32 @@
+{{- if $.Values.ingressInternal.enabled }}
+apiVersion: networking.k8s.io/v1
+kind: Ingress
+metadata:
+  annotations: {{ toYaml .Values.ingressInternal.annotations | nindent 4 }}
+  labels: {{ include "labels.standard" . | nindent 4 }}
+  name: {{ include "name" . }}-internal
+  namespace: {{ .Release.Namespace }}
+spec:
+  {{ if $.Values.ingressInternal.className }}
+  ingressClassName: {{ .Values.ingressInternal.className }}
+  {{ end }}
+  {{- with .Values.ingressInternal.tls }}
+  tls:
+    - hosts:
+        - {{ $.Values.domain | quote }}
+      {{- with .secretName }}
+      secretName: {{ . }}
+      {{- end }}
+  {{- end }}
+  rules:
+    - host: {{ .Values.domain }}
+      http:
+        paths:
+          - backend:
+              service:
+                name: {{ include "name" . }}
+                port:
+                  name: http
+            path: {{ $.Values.ingressInternal.path | default "/" }}
+            pathType: Prefix
+{{- end }}

chart/templates/ingress.yaml

@@ -0,0 +1,32 @@
+{{- if $.Values.ingress.enabled }}
+apiVersion: networking.k8s.io/v1
+kind: Ingress
+metadata:
+  annotations: {{ toYaml .Values.ingress.annotations | nindent 4 }}
+  labels: {{ include "labels.standard" . | nindent 4 }}
+  name: {{ include "name" . }}
+  namespace: {{ .Release.Namespace }}
+spec:
+  {{ if $.Values.ingress.className }}
+  ingressClassName: {{ .Values.ingress.className }}
+  {{ end }}
+  {{- with .Values.ingress.tls }}
+  tls:
+    - hosts:
+        - {{ $.Values.domain | quote }}
+      {{- with .secretName }}
+      secretName: {{ . }}
+      {{- end }}
+  {{- end }}
+  rules:
+    - host: {{ .Values.domain }}
+      http:
+        paths:
+          - backend:
+              service:
+                name: {{ include "name" . }}
+                port:
+                  name: http
+            path: {{ $.Values.ingress.path | default "/" }}
+            pathType: Prefix
+{{- end }}

chart/templates/network-policy.yaml

@@ -0,0 +1,36 @@
+{{- if $.Values.networkPolicy.enabled }}
+apiVersion: networking.k8s.io/v1
+kind: NetworkPolicy
+metadata:
+  name: {{ include "name" . }}
+  namespace: {{ .Release.Namespace }}
+spec:
+  egress:
+    - ports:
+        - port: 53
+          protocol: UDP
+      to:
+        - namespaceSelector:
+            matchLabels:
+              kubernetes.io/metadata.name: kube-system
+          podSelector:
+            matchLabels:
+              k8s-app: kube-dns
+    - to:
+        {{- range $ip := .Values.networkPolicy.allowedBlocks }}
+        - ipBlock:
+            cidr: {{ $ip | quote }}
+        {{- end }}
+    - to:
+        - ipBlock:
+            cidr: 0.0.0.0/0
+            except:
+              - 10.0.0.0/8
+              - 172.16.0.0/12
+              - 192.168.0.0/16
+              - 169.254.169.254/32
+  podSelector:
+    matchLabels: {{ include "labels.standard" . | nindent 6 }}
+  policyTypes:
+    - Egress
+{{- end }}

chart/templates/service-account.yaml

@@ -0,0 +1,13 @@
+{{- if and .Values.serviceAccount.enabled .Values.serviceAccount.create }}
+apiVersion: v1
+kind: ServiceAccount
+automountServiceAccountToken: {{ .Values.serviceAccount.automountServiceAccountToken }}
+metadata:
+  name: "{{ .Values.serviceAccount.name | default (include "name" .) }}"
+  namespace: {{ .Release.Namespace }}
+  labels: {{ include "labels.standard" . | nindent 4 }}
+  {{- with .Values.serviceAccount.annotations }}
+  annotations:
+    {{- toYaml . | nindent 4 }}
+  {{- end }}
+{{- end }}

chart/templates/service-monitor.yaml

@@ -0,0 +1,17 @@
+{{- if eq "true" $.Values.envVars.METRICS_ENABLED }}
+apiVersion: monitoring.coreos.com/v1
+kind: ServiceMonitor
+metadata:
+  labels: {{ include "labels.standard" . | nindent 4 }}
+  name: {{ include "name" . }}
+  namespace: {{ .Release.Namespace }}
+spec:
+  selector:
+    matchLabels: {{ include "labels.standard" . | nindent 6 }}
+  endpoints:
+    - port: metrics
+      path: /metrics
+      interval: 10s
+      scheme: http	
+      scrapeTimeout: 10s
+{{- end }}

chart/templates/service.yaml

@@ -0,0 +1,21 @@
+apiVersion: v1
+kind: Service
+metadata:
+  name: "{{ include "name" . }}"
+  annotations: {{ toYaml .Values.service.annotations | nindent 4 }}
+  namespace: {{ .Release.Namespace }}
+  labels: {{ include "labels.standard" . | nindent 4 }}
+spec:
+  ports:
+  - name: http
+    port: 80
+    protocol: TCP
+    targetPort: http
+  {{- if eq "true" $.Values.envVars.METRICS_ENABLED }}
+  - name: metrics
+    port: {{ $.Values.envVars.METRICS_PORT | default 5565 | int }}
+    protocol: TCP
+    targetPort: metrics
+  {{- end }}
+  selector: {{ include "labels.standard" . | nindent 4 }}
+  type: {{.Values.service.type}}

chart/values.yaml

@@ -0,0 +1,81 @@
+image:
+  repository: registry.internal.huggingface.tech/deepsite
+  name: deepsite
+  tag: 0.0.0-latest
+  pullPolicy: IfNotPresent
+
+replicas: 1
+
+domain: deepsite.hf.co
+
+networkPolicy:
+  enabled: true
+  allowedBlocks: []
+  # allowedBlocks:
+  #   - 10.0.240.0/24
+  #   - 10.0.241.0/24
+  #   - 10.0.242.0/24
+  #   - 10.0.243.0/24
+  #   - 10.0.244.0/24
+  #   - 10.240.0.0/24
+  #   - 10.16.0.0/16
+
+service:
+  type: NodePort
+  annotations: { }
+
+serviceAccount:
+  enabled: false
+  create: false
+  name: ""
+  automountServiceAccountToken: true
+  annotations: { }
+
+ingress:
+  enabled: true
+  path: "/"
+  annotations: { }
+  # className: "nginx"
+  tls: { }
+    # secretName: XXX
+
+ingressInternal:
+  enabled: false
+  path: "/"
+  annotations: { }
+  # className: "nginx"
+  tls: { }
+
+resources:
+  requests:
+    cpu: 2
+    memory: 4Gi
+  limits:
+    cpu: 2
+    memory: 4Gi
+nodeSelector: {}
+tolerations: []
+
+envVars: { }
+
+infisical:
+  enabled: false
+  env: ""
+  project: "deepsite-f-hvj"
+  url: ""
+  resyncInterval: 60
+  operatorSecretName: "deepsite-operator-secrets"
+  operatorSecretNamespace: "hub-utils"
+
+# Allow to environment injections on top or instead of infisical
+extraEnvFrom: []
+extraEnv: []
+
+autoscaling:
+  enabled: false
+  minReplicas: 1
+  maxReplicas: 2
+  targetMemoryUtilizationPercentage: ""
+  targetCPUUtilizationPercentage: ""
+
+## Metrics removed; monitoring configuration no longer used