feat(deploy): 增加三节点 K3s 高可用迁移能力

新增 WireGuard 全互联、三 server embedded-etcd K3s、CloudNativePG 双实例、Barman OSS 备份、双 NGINX、Kubernetes Secret/RBAC 与本地旧文件按严格 24 小时清理。

新增维护窗口数据迁移、digest 固定滚动发布、应用回滚、跨节点文件 E2E、节点和数据库故障演练、CNPG 恢复与 etcd 快照验收脚本；洛杉矶仅作为带 NoSchedule 污点的仲裁节点。

所有生产 Secret 只在执行时从 0600 本地环境和旧生产容器导入，仓库不保存凭据；公网入口保持人工 DNS 故障切换边界。

验证：bash -n、ShellCheck、kubectl kustomize、Node 语法检查、Secret 扫描、OSS put/head/delete 实测。

deploy/kubernetes/easyai-ai-gateway-cluster-release

@@ -0,0 +1,279 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+[[ $(id -u) -eq 0 ]] || { echo 'run as root' >&2; exit 1; }
+
+config_file=${AI_GATEWAY_CLUSTER_RELEASE_CONFIG:-/etc/easyai-ai-gateway-cluster-release.conf}
+manifest_tool=${AI_GATEWAY_RELEASE_MANIFEST_TOOL:-/usr/local/lib/easyai-ai-gateway-release/release-manifest.mjs}
+[[ -f $config_file && ! -L $config_file ]] || { echo 'missing cluster release config' >&2; exit 1; }
+[[ -f $manifest_tool && ! -L $manifest_tool ]] || { echo 'missing release manifest validator' >&2; exit 1; }
+# shellcheck source=/dev/null
+source "$config_file"
+
+: "${RELEASES_DIR:?}"
+: "${NAMESPACE:=easyai}"
+: "${PUBLIC_BASE_URL:=https://ai.51easyai.com}"
+: "${K3S_BIN:=/usr/local/bin/k3s}"
+[[ $RELEASES_DIR == /* && $NAMESPACE =~ ^[a-z0-9-]+$ ]] || {
+  echo 'invalid cluster release configuration' >&2
+  exit 1
+}
+[[ $PUBLIC_BASE_URL =~ ^https://[A-Za-z0-9.-]+$ ]] || {
+  echo 'invalid public base URL' >&2
+  exit 1
+}
+for command in "$K3S_BIN" curl flock install jq node; do
+  command -v "$command" >/dev/null 2>&1 || { echo "$command is required" >&2; exit 1; }
+done
+
+kubectl=("$K3S_BIN" kubectl)
+install -d -m 0700 "$RELEASES_DIR"
+exec 9>"$RELEASES_DIR/.release.lock"
+flock -n 9 || { echo 'another cluster release operation is running' >&2; exit 1; }
+
+manifest_get() {
+  node "$manifest_tool" get "$1" "$2"
+}
+
+current_to_file() {
+  local output=$1
+  "${kubectl[@]}" get configmap easyai-gateway-release -n "$NAMESPACE" \
+    -o jsonpath='{.data.manifest\.json}' >"$output"
+  [[ -s $output ]]
+  node "$manifest_tool" validate "$output" >/dev/null
+}
+
+wait_for_url() {
+  local label=$1
+  local url=$2
+  local expected=${3:-}
+  local body
+  for _ in $(seq 1 60); do
+    if body=$(curl -fsS --max-time 5 "$url" 2>/dev/null); then
+      if [[ -z $expected || $body == *"$expected"* ]]; then
+        echo "[cluster-release] verified $label"
+        return 0
+      fi
+    fi
+    sleep 2
+  done
+  echo "release probe failed: $label ($url)" >&2
+  return 1
+}
+
+verify_site() {
+  local site=$1
+  local address
+  case $site in
+    hongkong) address=10.77.0.2 ;;
+    ningbo) address=10.77.0.1 ;;
+    *) return 1 ;;
+  esac
+  wait_for_url "$site API health" "http://$address:30088/api/v1/healthz" easyai-ai-gateway
+  wait_for_url "$site API readiness" "http://$address:30088/api/v1/readyz" '"ok":true'
+  wait_for_url "$site Web" "http://$address:30178/" 'EasyAI AI Gateway'
+}
+
+rollout_site() {
+  local site=$1
+  local api_image=$2
+  local web_image=$3
+  local api_changed=$4
+  local web_changed=$5
+  if [[ $api_changed == true ]]; then
+    "${kubectl[@]}" set image "deployment/easyai-api-$site" -n "$NAMESPACE" \
+      "api=$api_image"
+    "${kubectl[@]}" rollout status "deployment/easyai-api-$site" -n "$NAMESPACE" --timeout=300s
+  fi
+  if [[ $web_changed == true ]]; then
+    "${kubectl[@]}" set image "deployment/easyai-web-$site" -n "$NAMESPACE" \
+      "web=$web_image"
+    "${kubectl[@]}" rollout status "deployment/easyai-web-$site" -n "$NAMESPACE" --timeout=300s
+  fi
+  verify_site "$site"
+}
+
+run_backup() {
+  local source_sha=$1
+  local backup=easyai-pre-release-${source_sha:0:12}
+  cat <<EOF | "${kubectl[@]}" apply -f - >/dev/null
+apiVersion: postgresql.cnpg.io/v1
+kind: Backup
+metadata:
+  name: $backup
+  namespace: $NAMESPACE
+spec:
+  cluster:
+    name: easyai-postgres
+  method: plugin
+  pluginConfiguration:
+    name: barman-cloud.cloudnative-pg.io
+EOF
+  for _ in $(seq 1 120); do
+    phase=$("${kubectl[@]}" get backup "$backup" -n "$NAMESPACE" \
+      -o jsonpath='{.status.phase}' 2>/dev/null || true)
+    [[ ${phase,,} == completed ]] && return 0
+    [[ ${phase,,} == failed ]] && break
+    sleep 5
+  done
+  echo "pre-migration backup failed: $backup phase=${phase:-unknown}" >&2
+  return 1
+}
+
+run_migrator() {
+  local source_sha=$1
+  local api_image=$2
+  local job=easyai-migrator-${source_sha:0:12}
+  "${kubectl[@]}" delete job "$job" -n "$NAMESPACE" \
+    --ignore-not-found=true --wait=true >/dev/null
+  cat <<EOF | "${kubectl[@]}" apply -f - >/dev/null
+apiVersion: batch/v1
+kind: Job
+metadata:
+  name: $job
+  namespace: $NAMESPACE
+  labels:
+    easyai.io/release: $source_sha
+spec:
+  backoffLimit: 0
+  ttlSecondsAfterFinished: 86400
+  template:
+    spec:
+      restartPolicy: Never
+      imagePullSecrets:
+        - name: easyai-registry
+      nodeSelector:
+        easyai.io/site: ningbo
+      containers:
+        - name: migrator
+          image: $api_image
+          command: ["/app/easyai-ai-gateway-migrate"]
+          envFrom:
+            - secretRef:
+                name: easyai-ai-gateway-runtime
+EOF
+  "${kubectl[@]}" wait --for=condition=complete "job/$job" \
+    -n "$NAMESPACE" --timeout=300s
+}
+
+record_current() {
+  local manifest=$1
+  local action=$2
+  local started_at=$3
+  local source_sha recorded temporary
+  source_sha=$(manifest_get "$manifest" sourceSha)
+  recorded=$RELEASES_DIR/$source_sha.json
+  temporary=$(mktemp "$RELEASES_DIR/.recorded.XXXXXX")
+  rm -f -- "$temporary"
+  RELEASE_DEPLOYED_AT=$(date -u '+%Y-%m-%dT%H:%M:%SZ') \
+  RELEASE_DEPLOY_DURATION_SECONDS=$(( $(date +%s) - started_at )) \
+  RELEASE_DEPLOY_ACTION=$action \
+    node "$manifest_tool" record-deployment "$manifest" "$temporary" >/dev/null
+  install -m 0600 "$temporary" "$recorded"
+  "${kubectl[@]}" create configmap easyai-gateway-release -n "$NAMESPACE" \
+    --from-file=manifest.json="$recorded" --dry-run=client -o yaml |
+    "${kubectl[@]}" apply -f - >/dev/null
+  rm -f -- "$temporary"
+}
+
+activate_manifest() {
+  local manifest=$1
+  local action=${2:-deploy}
+  local source_sha base_sha api_image web_image migrations_changed components
+  local api_changed=false web_changed=false current_file previous_api previous_web started_at
+  [[ -f $manifest && ! -L $manifest ]] || { echo 'manifest must be a regular file' >&2; return 1; }
+  node "$manifest_tool" validate "$manifest" >/dev/null
+  source_sha=$(manifest_get "$manifest" sourceSha)
+  base_sha=$(manifest_get "$manifest" baseReleaseSha)
+  api_image=$(manifest_get "$manifest" images.api)
+  web_image=$(manifest_get "$manifest" images.web)
+  migrations_changed=$(manifest_get "$manifest" migrationsChanged)
+  components=$(manifest_get "$manifest" components)
+  [[ $api_image =~ @sha256:[0-9a-f]{64}$ && $web_image =~ @sha256:[0-9a-f]{64}$ ]] || {
+    echo 'release images must be digest-pinned' >&2
+    return 1
+  }
+  [[ $components == *'"api"'* ]] && api_changed=true
+  [[ $components == *'"web"'* ]] && web_changed=true
+  if [[ $action == rollback ]]; then
+    api_changed=true
+    web_changed=true
+  fi
+
+  current_file=$(mktemp "$RELEASES_DIR/.current.XXXXXX")
+  if current_to_file "$current_file"; then
+    current_sha=$(manifest_get "$current_file" sourceSha)
+    previous_api=$(manifest_get "$current_file" images.api)
+    previous_web=$(manifest_get "$current_file" images.web)
+    if [[ $action == deploy && $current_sha != "$base_sha" ]]; then
+      echo "stale release manifest: production=$current_sha manifest_base=$base_sha" >&2
+      rm -f -- "$current_file"
+      return 1
+    fi
+  else
+    if [[ $action == deploy && -n $base_sha ]]; then
+      echo 'cluster release status is unavailable for a non-bootstrap manifest' >&2
+      rm -f -- "$current_file"
+      return 1
+    fi
+    previous_api=
+    previous_web=
+  fi
+  started_at=$(date +%s)
+
+  if [[ $action == deploy && $migrations_changed == true ]]; then
+    run_backup "$source_sha"
+    run_migrator "$source_sha" "$api_image"
+  fi
+
+  if ! rollout_site hongkong "$api_image" "$web_image" "$api_changed" "$web_changed" ||
+      ! rollout_site ningbo "$api_image" "$web_image" "$api_changed" "$web_changed" ||
+      ! wait_for_url 'public readiness' "$PUBLIC_BASE_URL/api/v1/readyz" '"ok":true'; then
+    if [[ -n $previous_api && -n $previous_web ]]; then
+      echo '[cluster-release] restoring previous application images' >&2
+      rollout_site hongkong "$previous_api" "$previous_web" "$api_changed" "$web_changed" || true
+      rollout_site ningbo "$previous_api" "$previous_web" "$api_changed" "$web_changed" || true
+    fi
+    rm -f -- "$current_file"
+    return 1
+  fi
+
+  record_current "$manifest" "$action" "$started_at"
+  rm -f -- "$current_file"
+  echo "production_release=PASS action=$action release=$source_sha"
+}
+
+case ${1:-} in
+  status)
+    [[ $# -eq 1 ]] || exit 64
+    status_file=$(mktemp "$RELEASES_DIR/.status.XXXXXX")
+    current_to_file "$status_file"
+    cat "$status_file"
+    rm -f -- "$status_file"
+    ;;
+  adopt)
+    [[ $# -eq 2 ]] || exit 64
+    node "$manifest_tool" validate "$2" >/dev/null
+    api_image=$(manifest_get "$2" images.api)
+    web_image=$(manifest_get "$2" images.web)
+    [[ $("${kubectl[@]}" get deployment easyai-api-ningbo -n "$NAMESPACE" \
+      -o jsonpath='{.spec.template.spec.containers[?(@.name=="api")].image}') == "$api_image" ]]
+    [[ $("${kubectl[@]}" get deployment easyai-web-ningbo -n "$NAMESPACE" \
+      -o jsonpath='{.spec.template.spec.containers[?(@.name=="web")].image}') == "$web_image" ]]
+    record_current "$2" deploy "$(date +%s)"
+    ;;
+  deploy)
+    [[ $# -eq 2 ]] || exit 64
+    activate_manifest "$2" deploy
+    ;;
+  rollback)
+    [[ $# -eq 2 && $2 =~ ^[0-9a-f]{40}$ ]] || exit 64
+    target=$RELEASES_DIR/$2.json
+    [[ -f $target && ! -L $target ]] || { echo 'unknown historical release' >&2; exit 1; }
+    activate_manifest "$target" rollback
+    ;;
+  *)
+    echo 'usage: easyai-ai-gateway-cluster-release {status|adopt <manifest>|deploy <manifest>|rollback <sha>}' >&2
+    exit 64
+    ;;
+esac

deploy/kubernetes/easyai-ai-gateway-cluster-release.conf.example

@@ -0,0 +1,4 @@
+RELEASES_DIR=/var/lib/easyai-ai-gateway-cluster-release/releases
+NAMESPACE=easyai
+PUBLIC_BASE_URL=https://ai.51easyai.com
+K3S_BIN=/usr/local/bin/k3s

deploy/kubernetes/production/application-config.yaml

@@ -0,0 +1,39 @@
+apiVersion: v1
+kind: ConfigMap
+metadata:
+  name: easyai-ai-gateway-config
+  labels:
+    app.kubernetes.io/part-of: easyai-ai-gateway
+data:
+  APP_ENV: production
+  HTTP_ADDR: ":8088"
+  IDENTITY_MODE: hybrid
+  IDENTITY_SECRET_STORE: kubernetes
+  IDENTITY_KUBERNETES_NAMESPACE: easyai
+  IDENTITY_KUBERNETES_SECRET_NAME: easyai-gateway-security-events
+  IDENTITY_SECURITY_EVENTS_HEARTBEAT_INTERVAL_SECONDS: "60"
+  IDENTITY_SECURITY_EVENTS_STALE_AFTER_SECONDS: "180"
+  IDENTITY_SECURITY_EVENTS_CLOCK_SKEW_SECONDS: "60"
+  SERVER_MAIN_BASE_URL: http://10.77.0.1:3000
+  TASK_PROGRESS_CALLBACK_ENABLED: "true"
+  TASK_PROGRESS_CALLBACK_URL: http://10.77.0.1:3000/internal/platform/task-progress-callbacks
+  TASK_PROGRESS_CALLBACK_TIMEOUT_MS: "5000"
+  TASK_PROGRESS_CALLBACK_MAX_ATTEMPTS: "10"
+  AI_GATEWAY_TASK_CLEANUP_ENABLED: "false"
+  AI_GATEWAY_TASK_RETENTION_DAYS: "30"
+  AI_GATEWAY_TASK_ANALYSIS_RETENTION_DAYS: "7"
+  AI_GATEWAY_TASK_CLEANUP_INTERVAL_SECONDS: "300"
+  AI_GATEWAY_TASK_CLEANUP_BATCH_SIZE: "1000"
+  CORS_ALLOWED_ORIGIN: https://ai.51easyai.com
+  AI_GATEWAY_PUBLIC_BASE_URL: https://ai.51easyai.com
+  AI_GATEWAY_WEB_BASE_URL: https://ai.51easyai.com
+  AI_GATEWAY_GENERATED_STORAGE_DIR: /app/data/static/generated
+  AI_GATEWAY_UPLOADED_STORAGE_DIR: /app/data/static/uploaded
+  AI_GATEWAY_LOCAL_TEMP_ASSET_TTL_HOURS: "24"
+  AI_GATEWAY_LOCAL_RESULT_TTL_HOURS: "24"
+  AI_GATEWAY_LOCAL_RESULT_MIN_FREE_BYTES: "10737418240"
+  AI_GATEWAY_LOCAL_RESULT_MAX_BYTES: "268435456"
+  AI_GATEWAY_LOCAL_RESULT_MAX_TASK_BYTES: "536870912"
+  AI_GATEWAY_ASYNC_QUEUE_WORKER_ENABLED: "true"
+  AI_GATEWAY_ASYNC_WORKER_HARD_LIMIT: "2048"
+  AI_GATEWAY_ASYNC_WORKER_REFRESH_INTERVAL_SECONDS: "5"

deploy/kubernetes/production/application.yaml

@@ -0,0 +1,439 @@
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: easyai-api-ningbo
+  labels:
+    app.kubernetes.io/name: easyai-api
+    app.kubernetes.io/component: api
+    app.kubernetes.io/part-of: easyai-ai-gateway
+spec:
+  replicas: 1
+  strategy:
+    type: RollingUpdate
+    rollingUpdate:
+      maxSurge: 1
+      maxUnavailable: 0
+  selector:
+    matchLabels:
+      app.kubernetes.io/name: easyai-api
+      easyai.io/site: ningbo
+  template:
+    metadata:
+      labels:
+        app.kubernetes.io/name: easyai-api
+        app.kubernetes.io/component: api
+        app.kubernetes.io/part-of: easyai-ai-gateway
+        easyai.io/site: ningbo
+    spec:
+      serviceAccountName: easyai-ai-gateway
+      terminationGracePeriodSeconds: 90
+      nodeSelector:
+        easyai.io/site: ningbo
+        easyai.io/workload: "true"
+      securityContext:
+        runAsNonRoot: true
+        seccompProfile:
+          type: RuntimeDefault
+      containers:
+        - name: api
+          image: easyai-api
+          imagePullPolicy: IfNotPresent
+          envFrom:
+            - configMapRef:
+                name: easyai-ai-gateway-config
+            - secretRef:
+                name: easyai-ai-gateway-runtime
+          env:
+            - name: POD_NAMESPACE
+              valueFrom:
+                fieldRef:
+                  fieldPath: metadata.namespace
+            - name: POD_NAME
+              valueFrom:
+                fieldRef:
+                  fieldPath: metadata.name
+          ports:
+            - name: http
+              containerPort: 8088
+          readinessProbe:
+            httpGet:
+              path: /api/v1/readyz
+              port: http
+            initialDelaySeconds: 3
+            periodSeconds: 5
+            timeoutSeconds: 3
+            failureThreshold: 6
+          livenessProbe:
+            httpGet:
+              path: /api/v1/healthz
+              port: http
+            initialDelaySeconds: 15
+            periodSeconds: 10
+            timeoutSeconds: 3
+            failureThreshold: 6
+          lifecycle:
+            preStop:
+              exec:
+                command: ["/bin/sh", "-c", "sleep 10"]
+          resources:
+            requests:
+              cpu: 250m
+              memory: 512Mi
+            limits:
+              cpu: "2"
+              memory: 2Gi
+          securityContext:
+            allowPrivilegeEscalation: false
+            capabilities:
+              drop: ["ALL"]
+            readOnlyRootFilesystem: true
+          volumeMounts:
+            - name: application-data
+              mountPath: /app/data
+            - name: temporary-files
+              mountPath: /tmp
+      volumes:
+        - name: application-data
+          emptyDir:
+            sizeLimit: 5Gi
+        - name: temporary-files
+          emptyDir:
+            sizeLimit: 1Gi
+---
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: easyai-api-hongkong
+  labels:
+    app.kubernetes.io/name: easyai-api
+    app.kubernetes.io/component: api
+    app.kubernetes.io/part-of: easyai-ai-gateway
+spec:
+  replicas: 1
+  strategy:
+    type: RollingUpdate
+    rollingUpdate:
+      maxSurge: 1
+      maxUnavailable: 0
+  selector:
+    matchLabels:
+      app.kubernetes.io/name: easyai-api
+      easyai.io/site: hongkong
+  template:
+    metadata:
+      labels:
+        app.kubernetes.io/name: easyai-api
+        app.kubernetes.io/component: api
+        app.kubernetes.io/part-of: easyai-ai-gateway
+        easyai.io/site: hongkong
+    spec:
+      serviceAccountName: easyai-ai-gateway
+      terminationGracePeriodSeconds: 90
+      nodeSelector:
+        easyai.io/site: hongkong
+        easyai.io/workload: "true"
+      securityContext:
+        runAsNonRoot: true
+        seccompProfile:
+          type: RuntimeDefault
+      containers:
+        - name: api
+          image: easyai-api
+          imagePullPolicy: IfNotPresent
+          envFrom:
+            - configMapRef:
+                name: easyai-ai-gateway-config
+            - secretRef:
+                name: easyai-ai-gateway-runtime
+          env:
+            - name: POD_NAMESPACE
+              valueFrom:
+                fieldRef:
+                  fieldPath: metadata.namespace
+            - name: POD_NAME
+              valueFrom:
+                fieldRef:
+                  fieldPath: metadata.name
+          ports:
+            - name: http
+              containerPort: 8088
+          readinessProbe:
+            httpGet:
+              path: /api/v1/readyz
+              port: http
+            initialDelaySeconds: 3
+            periodSeconds: 5
+            timeoutSeconds: 3
+            failureThreshold: 6
+          livenessProbe:
+            httpGet:
+              path: /api/v1/healthz
+              port: http
+            initialDelaySeconds: 15
+            periodSeconds: 10
+            timeoutSeconds: 3
+            failureThreshold: 6
+          lifecycle:
+            preStop:
+              exec:
+                command: ["/bin/sh", "-c", "sleep 10"]
+          resources:
+            requests:
+              cpu: 250m
+              memory: 512Mi
+            limits:
+              cpu: "2"
+              memory: 2Gi
+          securityContext:
+            allowPrivilegeEscalation: false
+            capabilities:
+              drop: ["ALL"]
+            readOnlyRootFilesystem: true
+          volumeMounts:
+            - name: application-data
+              mountPath: /app/data
+            - name: temporary-files
+              mountPath: /tmp
+      volumes:
+        - name: application-data
+          emptyDir:
+            sizeLimit: 5Gi
+        - name: temporary-files
+          emptyDir:
+            sizeLimit: 1Gi
+---
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: easyai-web-ningbo
+  labels:
+    app.kubernetes.io/name: easyai-web
+    app.kubernetes.io/component: web
+    app.kubernetes.io/part-of: easyai-ai-gateway
+spec:
+  replicas: 1
+  strategy:
+    type: RollingUpdate
+    rollingUpdate:
+      maxSurge: 1
+      maxUnavailable: 0
+  selector:
+    matchLabels:
+      app.kubernetes.io/name: easyai-web
+      easyai.io/site: ningbo
+  template:
+    metadata:
+      labels:
+        app.kubernetes.io/name: easyai-web
+        app.kubernetes.io/component: web
+        app.kubernetes.io/part-of: easyai-ai-gateway
+        easyai.io/site: ningbo
+    spec:
+      automountServiceAccountToken: false
+      terminationGracePeriodSeconds: 30
+      nodeSelector:
+        easyai.io/site: ningbo
+        easyai.io/workload: "true"
+      securityContext:
+        runAsNonRoot: true
+        runAsUser: 101
+        runAsGroup: 101
+        fsGroup: 101
+        fsGroupChangePolicy: OnRootMismatch
+        seccompProfile:
+          type: RuntimeDefault
+      containers:
+        - name: web
+          image: easyai-web
+          imagePullPolicy: IfNotPresent
+          ports:
+            - name: http
+              containerPort: 80
+          readinessProbe:
+            httpGet:
+              path: /
+              port: http
+            initialDelaySeconds: 3
+            periodSeconds: 5
+            timeoutSeconds: 3
+            failureThreshold: 6
+          livenessProbe:
+            httpGet:
+              path: /
+              port: http
+            initialDelaySeconds: 15
+            periodSeconds: 10
+            timeoutSeconds: 3
+            failureThreshold: 6
+          resources:
+            requests:
+              cpu: 50m
+              memory: 64Mi
+            limits:
+              cpu: 500m
+              memory: 256Mi
+          securityContext:
+            allowPrivilegeEscalation: false
+            capabilities:
+              drop: ["ALL"]
+              add: ["NET_BIND_SERVICE"]
+            readOnlyRootFilesystem: true
+          volumeMounts:
+            - name: nginx-cache
+              mountPath: /var/cache/nginx
+            - name: nginx-run
+              mountPath: /var/run
+      volumes:
+        - name: nginx-cache
+          emptyDir:
+            sizeLimit: 128Mi
+        - name: nginx-run
+          emptyDir:
+            sizeLimit: 16Mi
+---
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: easyai-web-hongkong
+  labels:
+    app.kubernetes.io/name: easyai-web
+    app.kubernetes.io/component: web
+    app.kubernetes.io/part-of: easyai-ai-gateway
+spec:
+  replicas: 1
+  strategy:
+    type: RollingUpdate
+    rollingUpdate:
+      maxSurge: 1
+      maxUnavailable: 0
+  selector:
+    matchLabels:
+      app.kubernetes.io/name: easyai-web
+      easyai.io/site: hongkong
+  template:
+    metadata:
+      labels:
+        app.kubernetes.io/name: easyai-web
+        app.kubernetes.io/component: web
+        app.kubernetes.io/part-of: easyai-ai-gateway
+        easyai.io/site: hongkong
+    spec:
+      automountServiceAccountToken: false
+      terminationGracePeriodSeconds: 30
+      nodeSelector:
+        easyai.io/site: hongkong
+        easyai.io/workload: "true"
+      securityContext:
+        runAsNonRoot: true
+        runAsUser: 101
+        runAsGroup: 101
+        fsGroup: 101
+        fsGroupChangePolicy: OnRootMismatch
+        seccompProfile:
+          type: RuntimeDefault
+      containers:
+        - name: web
+          image: easyai-web
+          imagePullPolicy: IfNotPresent
+          ports:
+            - name: http
+              containerPort: 80
+          readinessProbe:
+            httpGet:
+              path: /
+              port: http
+            initialDelaySeconds: 3
+            periodSeconds: 5
+            timeoutSeconds: 3
+            failureThreshold: 6
+          livenessProbe:
+            httpGet:
+              path: /
+              port: http
+            initialDelaySeconds: 15
+            periodSeconds: 10
+            timeoutSeconds: 3
+            failureThreshold: 6
+          resources:
+            requests:
+              cpu: 50m
+              memory: 64Mi
+            limits:
+              cpu: 500m
+              memory: 256Mi
+          securityContext:
+            allowPrivilegeEscalation: false
+            capabilities:
+              drop: ["ALL"]
+              add: ["NET_BIND_SERVICE"]
+            readOnlyRootFilesystem: true
+          volumeMounts:
+            - name: nginx-cache
+              mountPath: /var/cache/nginx
+            - name: nginx-run
+              mountPath: /var/run
+      volumes:
+        - name: nginx-cache
+          emptyDir:
+            sizeLimit: 128Mi
+        - name: nginx-run
+          emptyDir:
+            sizeLimit: 16Mi
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: api
+  labels:
+    app.kubernetes.io/name: easyai-api
+    app.kubernetes.io/component: api
+    app.kubernetes.io/part-of: easyai-ai-gateway
+spec:
+  type: NodePort
+  externalTrafficPolicy: Local
+  selector:
+    app.kubernetes.io/name: easyai-api
+  ports:
+    - name: http
+      port: 8088
+      targetPort: http
+      nodePort: 30088
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: web
+  labels:
+    app.kubernetes.io/name: easyai-web
+    app.kubernetes.io/component: web
+    app.kubernetes.io/part-of: easyai-ai-gateway
+spec:
+  type: NodePort
+  externalTrafficPolicy: Local
+  selector:
+    app.kubernetes.io/name: easyai-web
+  ports:
+    - name: http
+      port: 80
+      targetPort: http
+      nodePort: 30178
+---
+apiVersion: policy/v1
+kind: PodDisruptionBudget
+metadata:
+  name: easyai-api
+spec:
+  minAvailable: 1
+  selector:
+    matchLabels:
+      app.kubernetes.io/name: easyai-api
+---
+apiVersion: policy/v1
+kind: PodDisruptionBudget
+metadata:
+  name: easyai-web
+spec:
+  minAvailable: 1
+  selector:
+    matchLabels:
+      app.kubernetes.io/name: easyai-web

deploy/kubernetes/production/database.yaml

@@ -0,0 +1,113 @@
+apiVersion: barmancloud.cnpg.io/v1
+kind: ObjectStore
+metadata:
+  name: easyai-postgres-oss
+  labels:
+    app.kubernetes.io/part-of: easyai-ai-gateway
+spec:
+  retentionPolicy: 14d
+  configuration:
+    destinationPath: s3://wangbo0808/easyai-ai-gateway/production/postgresql
+    endpointURL: https://oss-cn-shanghai.aliyuncs.com
+    s3Credentials:
+      accessKeyId:
+        name: easyai-oss-backup
+        key: ACCESS_KEY_ID
+      secretAccessKey:
+        name: easyai-oss-backup
+        key: ACCESS_SECRET_KEY
+    wal:
+      compression: gzip
+      maxParallel: 2
+    data:
+      compression: gzip
+      jobs: 2
+  instanceSidecarConfiguration:
+    env:
+      - name: AWS_DEFAULT_REGION
+        value: oss-cn-shanghai
+      - name: AWS_REQUEST_CHECKSUM_CALCULATION
+        value: when_required
+      - name: AWS_RESPONSE_CHECKSUM_VALIDATION
+        value: when_required
+---
+apiVersion: postgresql.cnpg.io/v1
+kind: Cluster
+metadata:
+  name: easyai-postgres
+  labels:
+    app.kubernetes.io/name: easyai-postgres
+    app.kubernetes.io/part-of: easyai-ai-gateway
+spec:
+  instances: 2
+  imageName: ghcr.io/cloudnative-pg/postgresql:18.4-standard-trixie@sha256:4587df73024408f5b2be9b4dd6ba2ccee8c9e5dc0c9a87c274c292291cc8a68c
+  imagePullPolicy: IfNotPresent
+  primaryUpdateStrategy: unsupervised
+  primaryUpdateMethod: switchover
+  failoverDelay: 0
+  switchoverDelay: 60
+  stopDelay: 90
+  smartShutdownTimeout: 30
+  bootstrap:
+    initdb:
+      database: easyai_ai_gateway
+      owner: easyai
+      secret:
+        name: easyai-postgres-app
+      postInitSQL:
+        - CREATE EXTENSION IF NOT EXISTS pgcrypto;
+  storage:
+    storageClass: local-path
+    size: 20Gi
+  walStorage:
+    storageClass: local-path
+    size: 5Gi
+  affinity:
+    enablePodAntiAffinity: true
+    podAntiAffinityType: required
+    topologyKey: easyai.io/site
+    nodeSelector:
+      easyai.io/database: "true"
+  resources:
+    requests:
+      cpu: 500m
+      memory: 1Gi
+    limits:
+      cpu: "2"
+      memory: 3Gi
+  postgresql:
+    synchronous:
+      method: any
+      number: 1
+      dataDurability: preferred
+    parameters:
+      archive_timeout: 60s
+      max_connections: "200"
+      shared_buffers: 512MB
+      wal_compression: "on"
+      wal_keep_size: 1GB
+  monitoring:
+    enablePodMonitor: false
+  plugins:
+    - name: barman-cloud.cloudnative-pg.io
+      isWALArchiver: true
+      parameters:
+        barmanObjectName: easyai-postgres-oss
+---
+apiVersion: postgresql.cnpg.io/v1
+kind: ScheduledBackup
+metadata:
+  name: easyai-postgres-daily
+  labels:
+    app.kubernetes.io/part-of: easyai-ai-gateway
+spec:
+  # Six-field cron, UTC: 18:30 UTC is 02:30 Asia/Shanghai.
+  schedule: "0 30 18 * * *"
+  backupOwnerReference: self
+  cluster:
+    name: easyai-postgres
+  method: plugin
+  pluginConfiguration:
+    name: barman-cloud.cloudnative-pg.io
+  immediate: false
+  suspend: false

deploy/kubernetes/production/kustomization.yaml

@@ -0,0 +1,19 @@
+apiVersion: kustomize.config.k8s.io/v1beta1
+kind: Kustomization
+namespace: easyai
+resources:
+  - namespace.yaml
+  - service-account-rbac.yaml
+  - application-config.yaml
+  - application.yaml
+  - database.yaml
+  - network-policy.yaml
+images:
+  # Deployment tooling must replace both sentinel digests with release-manifest
+  # digests before server-side apply.
+  - name: easyai-api
+    newName: registry.cn-shanghai.aliyuncs.com/easyaigc/ai-gateway
+    digest: sha256:0000000000000000000000000000000000000000000000000000000000000000
+  - name: easyai-web
+    newName: registry.cn-shanghai.aliyuncs.com/easyaigc/ai-gateway-web
+    digest: sha256:0000000000000000000000000000000000000000000000000000000000000000

deploy/kubernetes/production/namespace.yaml

@@ -0,0 +1,9 @@
+apiVersion: v1
+kind: Namespace
+metadata:
+  name: easyai
+  labels:
+    app.kubernetes.io/part-of: easyai-ai-gateway
+    pod-security.kubernetes.io/enforce: baseline
+    pod-security.kubernetes.io/audit: restricted
+    pod-security.kubernetes.io/warn: restricted

deploy/kubernetes/production/network-policy.yaml

@@ -0,0 +1,23 @@
+apiVersion: networking.k8s.io/v1
+kind: NetworkPolicy
+metadata:
+  name: easyai-application-ingress
+spec:
+  podSelector:
+    matchExpressions:
+      - key: app.kubernetes.io/name
+        operator: In
+        values: [easyai-api, easyai-web]
+  policyTypes: [Ingress]
+  ingress:
+    - from:
+        - namespaceSelector:
+            matchLabels:
+              kubernetes.io/metadata.name: easyai
+        - ipBlock:
+            cidr: 10.77.0.0/24
+      ports:
+        - protocol: TCP
+          port: 8088
+        - protocol: TCP
+          port: 80

deploy/kubernetes/production/service-account-rbac.yaml

@@ -0,0 +1,31 @@
+apiVersion: v1
+kind: ServiceAccount
+metadata:
+  name: easyai-ai-gateway
+  labels:
+    app.kubernetes.io/part-of: easyai-ai-gateway
+automountServiceAccountToken: true
+imagePullSecrets:
+  - name: easyai-registry
+---
+apiVersion: rbac.authorization.k8s.io/v1
+kind: Role
+metadata:
+  name: easyai-gateway-security-events
+rules:
+  - apiGroups: [""]
+    resources: ["secrets"]
+    resourceNames: ["easyai-gateway-security-events"]
+    verbs: ["get", "update", "patch"]
+---
+apiVersion: rbac.authorization.k8s.io/v1
+kind: RoleBinding
+metadata:
+  name: easyai-gateway-security-events
+subjects:
+  - kind: ServiceAccount
+    name: easyai-ai-gateway
+roleRef:
+  apiGroup: rbac.authorization.k8s.io
+  kind: Role
+  name: easyai-gateway-security-events

deploy/nginx/ai.51easyai.com-cluster.conf

@@ -0,0 +1,106 @@
+upstream easyai_gateway_api {
+  least_conn;
+  keepalive 64;
+  server 10.77.0.1:30088 max_fails=2 fail_timeout=5s;
+  server 10.77.0.2:30088 max_fails=2 fail_timeout=5s;
+}
+
+upstream easyai_gateway_web {
+  least_conn;
+  keepalive 32;
+  server 10.77.0.1:30178 max_fails=2 fail_timeout=5s;
+  server 10.77.0.2:30178 max_fails=2 fail_timeout=5s;
+}
+
+server {
+  listen 80;
+  listen [::]:80;
+  server_name ai.51easyai.com;
+  return 301 https://$host$request_uri;
+}
+
+server {
+  listen 443 ssl http2;
+  listen [::]:443 ssl http2;
+  server_name ai.51easyai.com;
+
+  ssl_certificate /etc/nginx/tls/ai.51easyai.com/fullchain.pem;
+  ssl_certificate_key /etc/nginx/tls/ai.51easyai.com/privkey.pem;
+  ssl_session_cache shared:easyai_gateway_tls:10m;
+  ssl_session_timeout 1d;
+  ssl_protocols TLSv1.2 TLSv1.3;
+
+  client_max_body_size 200m;
+  client_body_timeout 300s;
+
+  include /etc/nginx/easyai-ai-gateway/legacy-static.inc;
+
+  location = /api/v1/metrics {
+    return 404;
+  }
+
+  location ^~ /api/v1/ {
+    proxy_http_version 1.1;
+    proxy_set_header Host $host;
+    proxy_set_header X-Real-IP $remote_addr;
+    proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
+    proxy_set_header X-Forwarded-Proto https;
+    proxy_set_header Connection "";
+    proxy_connect_timeout 2s;
+    proxy_send_timeout 3600s;
+    proxy_read_timeout 3600s;
+    proxy_request_buffering off;
+    proxy_buffering off;
+    proxy_next_upstream error timeout http_502 http_503 http_504;
+    proxy_next_upstream_tries 2;
+    proxy_pass http://easyai_gateway_api;
+  }
+
+  location /gateway-api/ {
+    proxy_http_version 1.1;
+    proxy_set_header Host $host;
+    proxy_set_header X-Real-IP $remote_addr;
+    proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
+    proxy_set_header X-Forwarded-Proto https;
+    proxy_set_header Connection "";
+    proxy_connect_timeout 2s;
+    proxy_send_timeout 3600s;
+    proxy_read_timeout 3600s;
+    proxy_request_buffering off;
+    proxy_buffering off;
+    proxy_next_upstream error timeout http_502 http_503 http_504;
+    proxy_next_upstream_tries 2;
+    proxy_pass http://easyai_gateway_api/;
+  }
+
+  location ^~ /static/ {
+    proxy_http_version 1.1;
+    proxy_set_header Host $host;
+    proxy_set_header X-Real-IP $remote_addr;
+    proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
+    proxy_set_header X-Forwarded-Proto https;
+    proxy_set_header Connection "";
+    proxy_connect_timeout 2s;
+    proxy_send_timeout 300s;
+    proxy_read_timeout 300s;
+    proxy_buffering off;
+    proxy_next_upstream error timeout http_502 http_503 http_504;
+    proxy_next_upstream_tries 2;
+    proxy_pass http://easyai_gateway_api;
+  }
+
+  location / {
+    proxy_http_version 1.1;
+    proxy_set_header Host $host;
+    proxy_set_header X-Real-IP $remote_addr;
+    proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
+    proxy_set_header X-Forwarded-Proto https;
+    proxy_set_header Connection "";
+    proxy_connect_timeout 2s;
+    proxy_send_timeout 300s;
+    proxy_read_timeout 300s;
+    proxy_next_upstream error timeout http_502 http_503 http_504;
+    proxy_next_upstream_tries 2;
+    proxy_pass http://easyai_gateway_web;
+  }
+}

deploy/nginx/ai.51easyai.com-maintenance.conf

@@ -0,0 +1,28 @@
+server {
+  listen 80;
+  listen [::]:80;
+  server_name ai.51easyai.com;
+  return 301 https://$host$request_uri;
+}
+
+server {
+  listen 443 ssl http2;
+  listen [::]:443 ssl http2;
+  server_name ai.51easyai.com;
+
+  ssl_certificate /etc/nginx/tls/ai.51easyai.com/fullchain.pem;
+  ssl_certificate_key /etc/nginx/tls/ai.51easyai.com/privkey.pem;
+  ssl_protocols TLSv1.2 TLSv1.3;
+
+  location ^~ /api/ {
+    default_type application/json;
+    add_header Retry-After 600 always;
+    return 503 '{"error":{"message":"Gateway database maintenance is in progress","type":"service_unavailable"}}';
+  }
+
+  location / {
+    default_type text/html;
+    add_header Retry-After 600 always;
+    return 503 '<!doctype html><html lang="zh-CN"><meta charset="utf-8"><title>系统维护</title><style>body{font-family:sans-serif;max-width:40rem;margin:15vh auto;padding:2rem;color:#1f2937}h1{font-size:1.8rem}</style><h1>系统正在进行高可用升级</h1><p>预计 5–10 分钟恢复，请稍后重试。</p></html>';
+  }
+}

deploy/nginx/cluster-acceptance-origin-ningbo.conf

@@ -0,0 +1,24 @@
+# Private-only deterministic path used by the cross-node acceptance test.
+server {
+  listen 10.77.0.1:18088;
+  server_name _;
+
+  allow 10.77.0.0/24;
+  deny all;
+  client_max_body_size 200m;
+
+  location / {
+    proxy_http_version 1.1;
+    proxy_set_header Host ai.51easyai.com;
+    proxy_set_header X-Real-IP $remote_addr;
+    proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
+    proxy_set_header X-Forwarded-Proto https;
+    proxy_set_header Connection "";
+    proxy_connect_timeout 2s;
+    proxy_send_timeout 3600s;
+    proxy_read_timeout 3600s;
+    proxy_request_buffering off;
+    proxy_buffering off;
+    proxy_pass http://10.77.0.1:30088;
+  }
+}

deploy/nginx/legacy-static-hongkong.inc

@@ -0,0 +1,20 @@
+# Hong Kong reads unexpired legacy files from the private Ningbo origin.
+location ^~ /static/generated/ {
+  proxy_http_version 1.1;
+  proxy_set_header Host ai.51easyai.com;
+  proxy_set_header Connection "";
+  proxy_connect_timeout 2s;
+  proxy_read_timeout 300s;
+  proxy_buffering off;
+  proxy_pass http://10.77.0.1:18080;
+}
+
+location ^~ /static/uploaded/ {
+  proxy_http_version 1.1;
+  proxy_set_header Host ai.51easyai.com;
+  proxy_set_header Connection "";
+  proxy_connect_timeout 2s;
+  proxy_read_timeout 300s;
+  proxy_buffering off;
+  proxy_pass http://10.77.0.1:18080;
+}

deploy/nginx/legacy-static-ningbo.inc

@@ -0,0 +1,13 @@
+# Temporary compatibility route. Remove after all files are strictly older
+# than 86,400 seconds and the cleanup acceptance check reports zero survivors.
+location ^~ /static/generated/ {
+  alias /var/lib/docker/volumes/easyai-ai-gateway_api_data/_data/static/generated/;
+  add_header Cache-Control "private, max-age=300";
+  try_files $request_filename =410;
+}
+
+location ^~ /static/uploaded/ {
+  alias /var/lib/docker/volumes/easyai-ai-gateway_api_data/_data/static/uploaded/;
+  add_header Cache-Control "private, max-age=300";
+  try_files $request_filename =410;
+}

deploy/nginx/legacy-static-origin-ningbo.conf

@@ -0,0 +1,19 @@
+server {
+  listen 10.77.0.1:18080;
+  server_name _;
+
+  allow 10.77.0.0/24;
+  deny all;
+
+  location ^~ /static/generated/ {
+    alias /var/lib/docker/volumes/easyai-ai-gateway_api_data/_data/static/generated/;
+    add_header Cache-Control "private, max-age=300";
+    try_files $request_filename =410;
+  }
+
+  location ^~ /static/uploaded/ {
+    alias /var/lib/docker/volumes/easyai-ai-gateway_api_data/_data/static/uploaded/;
+    add_header Cache-Control "private, max-age=300";
+    try_files $request_filename =410;
+  }
+}

deploy/systemd/easyai-legacy-generated-cleanup.service

@@ -0,0 +1,16 @@
+[Unit]
+Description=Prune EasyAI Gateway legacy local files older than 24 hours
+After=docker.service
+
+[Service]
+Type=oneshot
+ExecStart=/usr/local/sbin/easyai-prune-legacy-generated --apply
+User=root
+Group=root
+Nice=10
+IOSchedulingClass=idle
+NoNewPrivileges=true
+PrivateTmp=true
+ProtectHome=true
+ProtectSystem=strict
+ReadWritePaths=/var/lib/docker/volumes/easyai-ai-gateway_api_data/_data/static

deploy/systemd/easyai-legacy-generated-cleanup.timer

@@ -0,0 +1,11 @@
+[Unit]
+Description=Run EasyAI Gateway legacy local file cleanup hourly
+
+[Timer]
+OnCalendar=hourly
+Persistent=true
+RandomizedDelaySec=120
+Unit=easyai-legacy-generated-cleanup.service
+
+[Install]
+WantedBy=timers.target