diff --git a/k8s/DEPLOYMENT_SUMMARY.md b/k8s/DEPLOYMENT_SUMMARY.md
new file mode 100644
index 0000000..3a2c047
--- /dev/null
+++ b/k8s/DEPLOYMENT_SUMMARY.md
@@ -0,0 +1,129 @@
+# AKS 部署总结
+
+## ✅ 部署状态
+
+### 已成功部署的服务
+
+1. **MCP Server** ✅
+   - 状态: 运行中
+   - 健康检查: 通过
+   - 服务端点: `http://mcp-server:8000` (集群内)
+   - 端口转发: `kubectl port-forward svc/mcp-server 8002:8000 -n taiji-ai`
+
+2. **LiteLLM Gateway** ✅
+   - 状态: 运行中
+   - 健康检查: 通过（OpenRouter 账户需要充值）
+   - 服务端点: `http://litellm-gateway:4000` (集群内)
+   - 端口转发: `kubectl port-forward svc/litellm-gateway 4000:4000 -n taiji-ai`
+
+3. **Data Ingestion** ✅
+   - 状态: 运行中
+   - 服务端点: `http://data-ingestion:8000` (集群内)
+   - 端口转发: `kubectl port-forward svc/data-ingestion 8001:8000 -n taiji-ai`
+
+### 需要关注的服务
+
+1. **NATS** ⚠️
+   - 状态: 部分 Pod 在创建中
+   - 问题: 旧 Pod 使用 amd64 镜像导致失败，新 Pod 正在创建
+   - 建议: 等待新 Pod 完全启动后，删除旧的 CrashLoopBackOff Pod
+
+2. **Prometheus & Grafana** ⚠️
+   - 状态: PVC 挂载问题
+   - 建议: 检查存储类配置
+
+## 📊 部署信息
+
+- **AKS 集群**: taiji-ai-pda
+- **资源组**: taiji-ai-pda
+- **命名空间**: taiji-ai
+- **ACR**: taiji.azurecr.io
+- **节点架构**: ARM64
+
+## 🔧 测试命令
+
+### 健康检查
+
+```bash
+# MCP Server
+kubectl port-forward svc/mcp-server 8002:8000 -n taiji-ai
+curl http://localhost:8002/health
+
+# Data Ingestion
+kubectl port-forward svc/data-ingestion 8001:8000 -n taiji-ai
+curl http://localhost:8001/health
+
+# LiteLLM Gateway
+kubectl port-forward svc/litellm-gateway 4000:4000 -n taiji-ai
+curl -H "Authorization: Bearer sk-taiji-master-key" http://localhost:4000/health
+```
+
+### 查看日志
+
+```bash
+# MCP Server
+kubectl logs -f deployment/mcp-server -n taiji-ai
+
+# Data Ingestion
+kubectl logs -f deployment/data-ingestion -n taiji-ai
+
+# LiteLLM Gateway
+kubectl logs -f deployment/litellm-gateway -n taiji-ai
+```
+
+### 查看服务状态
+
+```bash
+kubectl get pods -n taiji-ai
+kubectl get services -n taiji-ai
+kubectl get deployments -n taiji-ai
+```
+
+## ⚠️ 已知问题
+
+1. **OpenRouter 账户余额不足**
+   - LiteLLM Gateway 健康检查显示模型端点不可用
+   - 需要访问 https://openrouter.ai/settings/credits 充值
+
+2. **NATS 集群配置**
+   - 当前配置为集群模式，但可能需要调整
+   - 如果不需要集群模式，可以简化为单节点部署
+
+3. **存储类**
+   - Prometheus 和 Grafana 的 PVC 可能需要检查存储类配置
+   - 当前使用 `managed-csi`，确保 AKS 集群支持
+
+## 🚀 下一步
+
+1. **清理旧 Pod**
+   ```bash
+   kubectl delete pod -l app=nats -n taiji-ai --field-selector=status.phase!=Running
+   ```
+
+2. **配置 Ingress**（可选）
+   - 如果需要外部访问，配置 Ingress
+   - 参考 `ingress.yaml` 文件
+
+3. **监控配置**
+   - 配置 Prometheus 和 Grafana
+   - 设置监控仪表板
+
+4. **OpenRouter 充值**
+   - 访问 https://openrouter.ai/settings/credits
+   - 充值后 LiteLLM Gateway 将正常工作
+
+## 📝 部署文件位置
+
+所有 Kubernetes 配置文件位于: `/home/taiji/tools/taiji-AI-PAD/k8s/`
+
+- `namespace.yaml` - 命名空间
+- `configmap.yaml` - 配置映射
+- `secrets.yaml.template` - Secret 模板
+- `*-deployment.yaml` - 各服务部署配置
+- `ingress.yaml` - Ingress 配置
+- `test-deployment.sh` - 测试脚本
+
+## 🎉 部署完成
+
+核心服务（MCP Server、LiteLLM Gateway、Data Ingestion）已成功部署并运行！
+
diff --git a/k8s/QUICKSTART.md b/k8s/QUICKSTART.md
new file mode 100644
index 0000000..ef69666
--- /dev/null
+++ b/k8s/QUICKSTART.md
@@ -0,0 +1,125 @@
+# AKS 快速部署指南
+
+## 🚀 5 步快速部署
+
+### 步骤 1: 准备 Azure 资源
+
+```bash
+# 设置变量（根据实际情况修改）
+export RESOURCE_GROUP="taiji-ai-rg"
+export LOCATION="southeastasia"
+export AKS_NAME="taiji-aks"
+export ACR_NAME="taijiacr"
+
+# 创建资源组
+az group create --name $RESOURCE_GROUP --location $LOCATION
+
+# 创建 ACR
+az acr create --resource-group $RESOURCE_GROUP --name $ACR_NAME --sku Basic
+
+# 创建 AKS 集群（最小配置）
+az aks create \
+  --resource-group $RESOURCE_GROUP \
+  --name $AKS_NAME \
+  --node-count 2 \
+  --node-vm-size Standard_B2s \
+  --enable-addons monitoring \
+  --attach-acr $ACR_NAME \
+  --generate-ssh-keys
+```
+
+### 步骤 2: 构建并推送镜像
+
+```bash
+cd k8s
+./build-and-push.sh $ACR_NAME
+```
+
+### 步骤 3: 创建 Secrets
+
+```bash
+# 使用交互式脚本
+./create-secrets.sh
+
+# 或手动创建
+kubectl create secret generic taiji-secrets \
+  --from-literal=database-url="postgresql+asyncpg://user:pass@server.postgres.database.azure.com:5432/db?sslmode=require" \
+  --from-literal=async-database-url="postgresql+asyncpg://user:pass@server.postgres.database.azure.com:5432/db?sslmode=require" \
+  --from-literal=redis-url="rediss://:password@server.redis.cache.windows.net:6380/0?ssl_cert_reqs=required" \
+  --from-literal=jwt-secret="your-secret-key" \
+  --from-literal=encryption-key="your-encryption-key" \
+  --from-literal=litellm-master-key="sk-taiji-master-key" \
+  --from-literal=litellm-api-key="sk-taiji-master-key" \
+  --from-literal=openrouter-api-key="your-openrouter-key" \
+  --from-literal=rapidapi-key="your-rapidapi-key" \
+  --namespace=taiji-ai
+```
+
+### 步骤 4: 部署到 AKS
+
+```bash
+./deploy.sh $ACR_NAME $RESOURCE_GROUP $AKS_NAME
+```
+
+### 步骤 5: 验证部署
+
+```bash
+# 查看所有 Pod 状态
+kubectl get pods -n taiji-ai
+
+# 查看服务
+kubectl get services -n taiji-ai
+
+# 查看日志
+kubectl logs -f deployment/mcp-server -n taiji-ai
+```
+
+## 📝 重要提示
+
+1. **数据库和 Redis**: 确保 Azure Database for PostgreSQL 和 Azure Cache for Redis 的防火墙规则允许 AKS 节点 IP 访问。
+
+2. **ACR 访问**: 如果 AKS 未自动附加 ACR，需要创建 ACR 拉取密钥：
+   ```bash
+   kubectl create secret docker-registry acr-secret \
+     --docker-server=${ACR_NAME}.azurecr.io \
+     --docker-username=<service-principal-id> \
+     --docker-password=<service-principal-password> \
+     --namespace=taiji-ai
+   ```
+
+3. **Ingress**: 默认配置使用 Azure Application Gateway。如果使用 NGINX Ingress，请修改 `ingress.yaml`。
+
+4. **域名**: 部署 Ingress 前，请确保域名 DNS 已正确配置。
+
+## 🔧 常用命令
+
+```bash
+# 查看所有资源
+kubectl get all -n taiji-ai
+
+# 扩展服务
+kubectl scale deployment/mcp-server --replicas=5 -n taiji-ai
+
+# 更新镜像
+kubectl set image deployment/mcp-server mcp-server=${ACR_NAME}.azurecr.io/mcp-server:latest -n taiji-ai
+
+# 进入 Pod
+kubectl exec -it <pod-name> -n taiji-ai -- /bin/bash
+
+# 查看事件
+kubectl get events -n taiji-ai --sort-by='.lastTimestamp'
+```
+
+## 🆘 故障排查
+
+如果 Pod 无法启动：
+
+1. 检查 Pod 状态: `kubectl describe pod <pod-name> -n taiji-ai`
+2. 查看日志: `kubectl logs <pod-name> -n taiji-ai`
+3. 检查 Secret: `kubectl get secret taiji-secrets -n taiji-ai -o yaml`
+4. 检查 ConfigMap: `kubectl get configmap taiji-config -n taiji-ai -o yaml`
+
+## 📚 更多信息
+
+详细文档请参考 [README.md](./README.md)
+
diff --git a/k8s/README.md b/k8s/README.md
new file mode 100644
index 0000000..6d168cd
--- /dev/null
+++ b/k8s/README.md
@@ -0,0 +1,328 @@
+# AKS 部署指南
+
+本文档说明如何将 taiji-AI-PAD 平台部署到 Azure Kubernetes Service (AKS)。
+
+## 📋 前置要求
+
+1. **Azure 账户和订阅**
+2. **已安装的工具**:
+   - Azure CLI (`az`)
+   - kubectl
+   - Docker
+   - envsubst (通常包含在 gettext 包中)
+
+3. **Azure 资源**:
+   - AKS 集群
+   - Azure Container Registry (ACR)
+   - Azure Database for PostgreSQL
+   - Azure Cache for Redis
+
+## 🚀 快速开始
+
+### 1. 准备 Azure 资源
+
+#### 创建 AKS 集群
+
+```bash
+# 设置变量
+RESOURCE_GROUP="taiji-ai-rg"
+LOCATION="southeastasia"
+AKS_NAME="taiji-aks"
+ACR_NAME="taijiacr"
+
+# 创建资源组
+az group create --name $RESOURCE_GROUP --location $LOCATION
+
+# 创建 ACR
+az acr create --resource-group $RESOURCE_GROUP --name $ACR_NAME --sku Basic
+
+# 创建 AKS 集群
+az aks create \
+  --resource-group $RESOURCE_GROUP \
+  --name $AKS_NAME \
+  --node-count 3 \
+  --enable-addons monitoring \
+  --attach-acr $ACR_NAME \
+  --generate-ssh-keys
+```
+
+#### 配置 Azure 数据库和 Redis
+
+确保您已经创建了：
+- Azure Database for PostgreSQL
+- Azure Cache for Redis
+
+记录连接字符串，稍后需要配置到 Secret 中。
+
+### 2. 构建和推送 Docker 镜像
+
+```bash
+# 给脚本添加执行权限
+chmod +x build-and-push.sh
+
+# 构建并推送所有镜像
+./build-and-push.sh $ACR_NAME
+```
+
+这将构建并推送以下镜像：
+- `litellm-gateway:latest`
+- `data-ingestion:latest`
+- `mcp-server:latest`
+
+### 3. 创建 Kubernetes Secrets
+
+#### 方式1: 使用交互式脚本
+
+```bash
+chmod +x create-secrets.sh
+./create-secrets.sh
+```
+
+#### 方式2: 使用 kubectl 命令
+
+```bash
+kubectl create secret generic taiji-secrets \
+  --from-literal=database-url="postgresql+asyncpg://..." \
+  --from-literal=redis-url="rediss://..." \
+  --from-literal=jwt-secret="your-jwt-secret" \
+  --from-literal=encryption-key="your-encryption-key" \
+  --from-literal=litellm-master-key="sk-taiji-master-key" \
+  --from-literal=openrouter-api-key="your-openrouter-key" \
+  --namespace=taiji-ai
+```
+
+#### 方式3: 使用 Azure Key Vault (推荐生产环境)
+
+1. 安装 Azure Key Vault Provider:
+```bash
+kubectl apply -f https://raw.githubusercontent.com/Azure/secrets-store-csi-driver-provider-azure/master/deployment/secrets-store-csi-driver.yaml
+```
+
+2. 创建 Key Vault 并存储密钥
+3. 配置 SecretProviderClass (参考 Azure 文档)
+
+### 4. 部署到 AKS
+
+```bash
+# 给脚本添加执行权限
+chmod +x deploy.sh
+
+# 部署所有服务
+./deploy.sh $ACR_NAME $RESOURCE_GROUP $AKS_NAME
+```
+
+## 📁 文件结构
+
+```
+k8s/
+├── namespace.yaml                    # 命名空间定义
+├── configmap.yaml                    # 非敏感配置
+├── secrets.yaml.template             # Secret 模板（参考用）
+├── nats-deployment.yaml              # NATS 消息队列部署
+├── model-gateway-deployment.yaml     # LiteLLM 网关部署
+├── data-ingestion-deployment.yaml    # 数据接入服务部署
+├── mcp-server-deployment.yaml        # MCP 服务器部署
+├── monitoring-deployment.yaml        # Prometheus 和 Grafana
+├── ingress.yaml                      # Ingress 配置
+├── deploy.sh                         # 部署脚本
+├── build-and-push.sh                # 镜像构建脚本
+├── create-secrets.sh                # Secret 创建脚本
+└── README.md                         # 本文档
+```
+
+## 🔧 配置说明
+
+### 环境变量
+
+所有敏感信息存储在 Kubernetes Secret 中，非敏感配置存储在 ConfigMap 中。
+
+**Secret 中的配置**:
+- `database-url`: PostgreSQL 连接字符串
+- `redis-url`: Redis 连接字符串
+- `jwt-secret`: JWT 签名密钥
+- `encryption-key`: 数据加密密钥
+- `litellm-master-key`: LiteLLM 主密钥
+- `openrouter-api-key`: OpenRouter API 密钥
+- `rapidapi-key`: RapidAPI 密钥（可选）
+- `azure-storage-connection-string`: Azure 存储连接字符串（可选）
+
+**ConfigMap 中的配置**:
+- NATS URL
+- LiteLLM URL
+- OpenRouter Base URL
+- 应用环境配置
+
+### 资源限制
+
+每个服务的默认资源限制：
+
+| 服务 | CPU 请求 | CPU 限制 | 内存请求 | 内存限制 |
+|------|---------|----------|----------|----------|
+| NATS | 100m | 500m | 256Mi | 512Mi |
+| Model Gateway | 250m | 1000m | 512Mi | 1Gi |
+| Data Ingestion | 250m | 1000m | 512Mi | 1Gi |
+| MCP Server | 250m | 500m | 512Mi | 1Gi |
+
+### 自动扩缩容 (HPA)
+
+所有主要服务都配置了 HorizontalPodAutoscaler：
+- **最小副本数**: 2-3
+- **最大副本数**: 8-10
+- **CPU 阈值**: 70%
+- **内存阈值**: 80%
+
+## 🌐 Ingress 配置
+
+### 选项1: Azure Application Gateway (推荐)
+
+如果使用 Azure Application Gateway Ingress Controller，配置文件已包含相应注解。
+
+### 选项2: NGINX Ingress
+
+如果使用 NGINX Ingress Controller，请：
+1. 安装 NGINX Ingress Controller
+2. 在 `ingress.yaml` 中取消注释 NGINX 配置
+3. 注释掉 AGIC 配置
+
+### 域名配置
+
+默认配置使用以下域名：
+- `api.taiji-ai.com` - MCP Server 和 Data Ingestion API
+- `gateway.taiji-ai.com` - Model Gateway (可选)
+
+请根据实际情况修改域名和 TLS 证书配置。
+
+## 📊 监控
+
+### Prometheus
+
+Prometheus 自动收集以下服务的指标：
+- LiteLLM Gateway
+- MCP Server
+- Data Ingestion
+
+访问方式：
+```bash
+kubectl port-forward svc/prometheus 9090:9090 -n taiji-ai
+# 然后访问 http://localhost:9090
+```
+
+### Grafana
+
+Grafana 已配置 Prometheus 作为数据源。
+
+访问方式：
+```bash
+kubectl port-forward svc/grafana 3000:3000 -n taiji-ai
+# 然后访问 http://localhost:3000
+# 默认用户名: admin
+# 密码: 在 Secret 中配置
+```
+
+## 🔍 故障排查
+
+### 查看 Pod 状态
+
+```bash
+kubectl get pods -n taiji-ai
+```
+
+### 查看 Pod 日志
+
+```bash
+# MCP Server
+kubectl logs -f deployment/mcp-server -n taiji-ai
+
+# Model Gateway
+kubectl logs -f deployment/litellm-gateway -n taiji-ai
+
+# Data Ingestion
+kubectl logs -f deployment/data-ingestion -n taiji-ai
+```
+
+### 查看服务状态
+
+```bash
+kubectl get services -n taiji-ai
+kubectl get ingress -n taiji-ai
+```
+
+### 检查 Secret 和 ConfigMap
+
+```bash
+kubectl get secrets -n taiji-ai
+kubectl get configmap -n taiji-ai
+kubectl describe secret taiji-secrets -n taiji-ai
+```
+
+### 常见问题
+
+1. **镜像拉取失败**
+   - 检查 ACR 是否正确附加到 AKS
+   - 验证 `acr-secret` 是否存在
+
+2. **Pod 无法启动**
+   - 检查 Secret 是否包含所有必需的键
+   - 查看 Pod 事件: `kubectl describe pod <pod-name> -n taiji-ai`
+
+3. **服务无法访问**
+   - 检查 Service 和 Ingress 配置
+   - 验证域名 DNS 配置
+
+4. **数据库连接失败**
+   - 检查 Azure Database for PostgreSQL 防火墙规则
+   - 验证连接字符串格式
+
+## 🔄 更新部署
+
+### 更新镜像
+
+```bash
+# 1. 构建新镜像
+./build-and-push.sh $ACR_NAME
+
+# 2. 重启部署（触发拉取新镜像）
+kubectl rollout restart deployment/mcp-server -n taiji-ai
+kubectl rollout restart deployment/litellm-gateway -n taiji-ai
+kubectl rollout restart deployment/data-ingestion -n taiji-ai
+```
+
+### 更新配置
+
+```bash
+# 更新 ConfigMap
+kubectl apply -f configmap.yaml
+
+# 更新 Secret
+kubectl create secret generic taiji-secrets \
+  --from-literal=... \
+  --namespace=taiji-ai \
+  --dry-run=client -o yaml | kubectl apply -f -
+
+# 重启相关 Pod
+kubectl rollout restart deployment/<service-name> -n taiji-ai
+```
+
+## 🧹 清理
+
+删除所有资源：
+
+```bash
+kubectl delete namespace taiji-ai
+```
+
+## 📚 参考资源
+
+- [AKS 文档](https://docs.microsoft.com/azure/aks/)
+- [Azure Container Registry 文档](https://docs.microsoft.com/azure/container-registry/)
+- [Kubernetes 文档](https://kubernetes.io/docs/)
+- [Azure Key Vault Provider](https://azure.github.io/secrets-store-csi-driver-provider-azure/)
+
+## 🆘 获取帮助
+
+如遇问题，请检查：
+1. Pod 日志
+2. 服务事件
+3. Ingress 状态
+4. Azure 资源健康状态
+
diff --git a/k8s/build-and-push.sh b/k8s/build-and-push.sh
new file mode 100755
index 0000000..15db473
--- /dev/null
+++ b/k8s/build-and-push.sh
@@ -0,0 +1,94 @@
+#!/bin/bash
+
+# 构建并推送 Docker 镜像到 ACR 的脚本
+# 使用方法: ./build-and-push.sh <ACR_NAME>
+
+set -e
+
+# 颜色输出
+RED='\033[0;31m'
+GREEN='\033[0;32m'
+YELLOW='\033[1;33m'
+NC='\033[0m'
+
+# 检查参数
+if [ $# -lt 1 ]; then
+    echo -e "${RED}错误: 缺少参数${NC}"
+    echo "使用方法: $0 <ACR_NAME>"
+    echo "示例: $0 myregistry"
+    exit 1
+fi
+
+ACR_NAME=$1
+ACR_URL="${ACR_NAME}.azurecr.io"
+
+echo -e "${GREEN}开始构建和推送镜像到 ACR...${NC}"
+echo "ACR 名称: $ACR_NAME"
+echo "ACR URL: $ACR_URL"
+
+# 检查 Docker 是否安装
+if ! command -v docker &> /dev/null; then
+    echo -e "${RED}错误: Docker 未安装${NC}"
+    exit 1
+fi
+
+# 检查 az 是否安装
+if ! command -v az &> /dev/null; then
+    echo -e "${RED}错误: Azure CLI 未安装${NC}"
+    exit 1
+fi
+
+# 登录到 ACR
+echo -e "${YELLOW}登录到 ACR...${NC}"
+az acr login --name $ACR_NAME
+
+# 获取项目根目录
+SCRIPT_DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" && pwd )"
+PROJECT_ROOT="$(dirname "$SCRIPT_DIR")"
+
+# 设置构建平台（ARM64）
+PLATFORM="linux/arm64"
+
+# 检查并设置 buildx
+if ! docker buildx ls | grep -q multiarch-builder; then
+    echo -e "${YELLOW}创建 buildx builder...${NC}"
+    docker buildx create --name multiarch-builder --use --bootstrap 2>/dev/null || true
+fi
+
+# 构建和推送 Model Gateway
+echo -e "${YELLOW}构建 Model Gateway 镜像 (${PLATFORM})...${NC}"
+cd "$PROJECT_ROOT/services/model-gateway"
+docker buildx build --platform ${PLATFORM} -t ${ACR_URL}/litellm-gateway:latest --push . || {
+    echo -e "${YELLOW}尝试使用标准构建...${NC}"
+    docker build -t ${ACR_URL}/litellm-gateway:latest .
+    docker push ${ACR_URL}/litellm-gateway:latest
+}
+echo -e "${GREEN}Model Gateway 镜像已推送${NC}"
+
+# 构建和推送 Data Ingestion
+echo -e "${YELLOW}构建 Data Ingestion 镜像 (${PLATFORM})...${NC}"
+cd "$PROJECT_ROOT/services/data-ingestion"
+docker buildx build --platform ${PLATFORM} -t ${ACR_URL}/data-ingestion:latest --push . || {
+    echo -e "${YELLOW}尝试使用标准构建...${NC}"
+    docker build -t ${ACR_URL}/data-ingestion:latest .
+    docker push ${ACR_URL}/data-ingestion:latest
+}
+echo -e "${GREEN}Data Ingestion 镜像已推送${NC}"
+
+# 构建和推送 MCP Server
+echo -e "${YELLOW}构建 MCP Server 镜像 (${PLATFORM})...${NC}"
+cd "$PROJECT_ROOT/services/mcp-server"
+docker buildx build --platform ${PLATFORM} -t ${ACR_URL}/mcp-server:latest --push . || {
+    echo -e "${YELLOW}尝试使用标准构建...${NC}"
+    docker build -t ${ACR_URL}/mcp-server:latest .
+    docker push ${ACR_URL}/mcp-server:latest
+}
+echo -e "${GREEN}MCP Server 镜像已推送${NC}"
+
+echo -e "${GREEN}所有镜像构建和推送完成！${NC}"
+echo ""
+echo "镜像列表:"
+echo "  - ${ACR_URL}/litellm-gateway:latest"
+echo "  - ${ACR_URL}/data-ingestion:latest"
+echo "  - ${ACR_URL}/mcp-server:latest"
+
diff --git a/k8s/configmap.yaml b/k8s/configmap.yaml
new file mode 100644
index 0000000..c26dbb8
--- /dev/null
+++ b/k8s/configmap.yaml
@@ -0,0 +1,34 @@
+# ConfigMap - 非敏感配置
+apiVersion: v1
+kind: ConfigMap
+metadata:
+  name: taiji-config
+  namespace: taiji-ai
+data:
+  # NATS配置
+  nats-url: "nats://nats:4222"
+  
+  # LiteLLM配置
+  litellm-url: "http://litellm-gateway:4000"
+  litellm-config-path: "/app/config/litellm_simple.yaml"
+  
+  # OpenRouter配置
+  openrouter-base-url: "https://openrouter.ai/api/v1"
+  
+  # RapidAPI配置
+  rapidapi-host: "rapidapi.com"
+  rapidapi-base-url: "https://rapidapi.com"
+  
+  # 应用配置
+  environment: "production"
+  debug: "false"
+  log-level: "INFO"
+  
+  # 健康检查配置
+  health-check-interval: "30"
+  
+  # 资源限制
+  max-agents-per-user: "100"
+  agent-execution-timeout: "300"
+  tool-execution-timeout: "60"
+
diff --git a/k8s/create-secrets.sh b/k8s/create-secrets.sh
new file mode 100755
index 0000000..1304480
--- /dev/null
+++ b/k8s/create-secrets.sh
@@ -0,0 +1,110 @@
+#!/bin/bash
+
+# 创建 Kubernetes Secrets 的交互式脚本
+# 使用方法: ./create-secrets.sh
+
+set -e
+
+# 颜色输出
+RED='\033[0;31m'
+GREEN='\033[0;32m'
+YELLOW='\033[1;33m'
+NC='\033[0m'
+
+echo -e "${GREEN}创建 Kubernetes Secrets${NC}"
+echo ""
+
+# 检查 kubectl
+if ! command -v kubectl &> /dev/null; then
+    echo -e "${RED}错误: kubectl 未安装${NC}"
+    exit 1
+fi
+
+# 检查命名空间
+NAMESPACE="taiji-ai"
+if ! kubectl get namespace $NAMESPACE &> /dev/null; then
+    echo -e "${YELLOW}创建命名空间 $NAMESPACE...${NC}"
+    kubectl create namespace $NAMESPACE
+fi
+
+# 提示输入各个 Secret 值
+echo -e "${YELLOW}请输入以下配置信息（按 Enter 跳过可选项）:${NC}"
+echo ""
+
+read -p "数据库 URL (DATABASE_URL): " DATABASE_URL
+read -p "异步数据库 URL (ASYNC_DATABASE_URL): " ASYNC_DATABASE_URL
+read -p "Redis URL (REDIS_URL): " REDIS_URL
+read -p "JWT Secret: " JWT_SECRET
+read -p "加密密钥 (ENCRYPTION_KEY): " ENCRYPTION_KEY
+read -p "LiteLLM Master Key: " LITELLM_MASTER_KEY
+read -p "OpenRouter API Key: " OPENROUTER_API_KEY
+read -p "RapidAPI Key (可选): " RAPIDAPI_KEY
+read -p "Azure Storage Connection String (可选): " AZURE_STORAGE_CONNECTION
+
+# 构建 kubectl 命令
+SECRET_ARGS=()
+
+if [ -n "$DATABASE_URL" ]; then
+    SECRET_ARGS+=("--from-literal=database-url=$DATABASE_URL")
+fi
+
+if [ -n "$ASYNC_DATABASE_URL" ]; then
+    SECRET_ARGS+=("--from-literal=async-database-url=$ASYNC_DATABASE_URL")
+fi
+
+if [ -n "$REDIS_URL" ]; then
+    SECRET_ARGS+=("--from-literal=redis-url=$REDIS_URL")
+fi
+
+if [ -n "$JWT_SECRET" ]; then
+    SECRET_ARGS+=("--from-literal=jwt-secret=$JWT_SECRET")
+fi
+
+if [ -n "$ENCRYPTION_KEY" ]; then
+    SECRET_ARGS+=("--from-literal=encryption-key=$ENCRYPTION_KEY")
+fi
+
+if [ -n "$LITELLM_MASTER_KEY" ]; then
+    SECRET_ARGS+=("--from-literal=litellm-master-key=$LITELLM_MASTER_KEY")
+    SECRET_ARGS+=("--from-literal=litellm-api-key=$LITELLM_MASTER_KEY")
+fi
+
+if [ -n "$OPENROUTER_API_KEY" ]; then
+    SECRET_ARGS+=("--from-literal=openrouter-api-key=$OPENROUTER_API_KEY")
+fi
+
+if [ -n "$RAPIDAPI_KEY" ]; then
+    SECRET_ARGS+=("--from-literal=rapidapi-key=$RAPIDAPI_KEY")
+fi
+
+if [ -n "$AZURE_STORAGE_CONNECTION" ]; then
+    SECRET_ARGS+=("--from-literal=azure-storage-connection-string=$AZURE_STORAGE_CONNECTION")
+fi
+
+# 删除现有 Secret（如果存在）
+if kubectl get secret taiji-secrets -n $NAMESPACE &> /dev/null; then
+    echo -e "${YELLOW}删除现有 Secret...${NC}"
+    kubectl delete secret taiji-secrets -n $NAMESPACE
+fi
+
+# 创建 Secret
+if [ ${#SECRET_ARGS[@]} -gt 0 ]; then
+    echo -e "${YELLOW}创建 Secret...${NC}"
+    kubectl create secret generic taiji-secrets \
+        "${SECRET_ARGS[@]}" \
+        --namespace=$NAMESPACE
+    
+    echo -e "${GREEN}Secret 创建成功！${NC}"
+else
+    echo -e "${RED}错误: 没有提供任何 Secret 值${NC}"
+    exit 1
+fi
+
+# 验证 Secret
+echo ""
+echo -e "${YELLOW}验证 Secret:${NC}"
+kubectl get secret taiji-secrets -n $NAMESPACE
+
+echo ""
+echo -e "${GREEN}完成！${NC}"
+
diff --git a/k8s/data-ingestion-deployment.yaml b/k8s/data-ingestion-deployment.yaml
new file mode 100644
index 0000000..9e65cd1
--- /dev/null
+++ b/k8s/data-ingestion-deployment.yaml
@@ -0,0 +1,153 @@
+# Data Ingestion 服务部署配置
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: data-ingestion
+  namespace: taiji-ai
+  labels:
+    app: data-ingestion
+    component: ingestion
+spec:
+  replicas: 2
+  selector:
+    matchLabels:
+      app: data-ingestion
+  template:
+    metadata:
+      labels:
+        app: data-ingestion
+        component: ingestion
+    spec:
+      containers:
+      - name: data-ingestion
+        image: ${ACR_NAME}.azurecr.io/data-ingestion:latest
+        imagePullPolicy: Always
+        ports:
+        - containerPort: 8000
+          name: http
+          protocol: TCP
+        env:
+        - name: DATABASE_URL
+          valueFrom:
+            secretKeyRef:
+              name: taiji-secrets
+              key: database-url
+        - name: ASYNC_DATABASE_URL
+          valueFrom:
+            secretKeyRef:
+              name: taiji-secrets
+              key: async-database-url
+        - name: REDIS_URL
+          valueFrom:
+            secretKeyRef:
+              name: taiji-secrets
+              key: redis-url
+        - name: NATS_URL
+          valueFrom:
+            configMapKeyRef:
+              name: taiji-config
+              key: nats-url
+        - name: RAPIDAPI_KEY
+          valueFrom:
+            secretKeyRef:
+              name: taiji-secrets
+              key: rapidapi-key
+        - name: RAPIDAPI_HOST
+          valueFrom:
+            configMapKeyRef:
+              name: taiji-config
+              key: rapidapi-host
+        - name: OPENROUTER_API_KEY
+          valueFrom:
+            secretKeyRef:
+              name: taiji-secrets
+              key: openrouter-api-key
+        - name: OPENROUTER_BASE_URL
+          valueFrom:
+            configMapKeyRef:
+              name: taiji-config
+              key: openrouter-base-url
+        - name: PYTHONUNBUFFERED
+          value: "1"
+        resources:
+          requests:
+            memory: "512Mi"
+            cpu: "250m"
+          limits:
+            memory: "1Gi"
+            cpu: "1000m"
+        volumeMounts:
+        - name: logs
+          mountPath: /app/logs
+        - name: cache
+          mountPath: /app/cache
+        livenessProbe:
+          httpGet:
+            path: /health
+            port: 8000
+          initialDelaySeconds: 30
+          periodSeconds: 10
+          timeoutSeconds: 5
+          failureThreshold: 3
+        readinessProbe:
+          httpGet:
+            path: /health
+            port: 8000
+          initialDelaySeconds: 10
+          periodSeconds: 5
+          timeoutSeconds: 3
+          failureThreshold: 2
+      volumes:
+      - name: logs
+        emptyDir: {}
+      - name: cache
+        emptyDir: {}
+      imagePullSecrets:
+      - name: acr-secret
+
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: data-ingestion
+  namespace: taiji-ai
+  labels:
+    app: data-ingestion
+spec:
+  type: ClusterIP
+  ports:
+  - port: 8000
+    targetPort: 8000
+    protocol: TCP
+    name: http
+  selector:
+    app: data-ingestion
+
+---
+# HorizontalPodAutoscaler
+apiVersion: autoscaling/v2
+kind: HorizontalPodAutoscaler
+metadata:
+  name: data-ingestion-hpa
+  namespace: taiji-ai
+spec:
+  scaleTargetRef:
+    apiVersion: apps/v1
+    kind: Deployment
+    name: data-ingestion
+  minReplicas: 2
+  maxReplicas: 8
+  metrics:
+  - type: Resource
+    resource:
+      name: cpu
+      target:
+        type: Utilization
+        averageUtilization: 70
+  - type: Resource
+    resource:
+      name: memory
+      target:
+        type: Utilization
+        averageUtilization: 80
+
diff --git a/k8s/deploy.sh b/k8s/deploy.sh
new file mode 100755
index 0000000..48ebe9b
--- /dev/null
+++ b/k8s/deploy.sh
@@ -0,0 +1,130 @@
+#!/bin/bash
+
+# AKS 部署脚本
+# 使用方法: ./deploy.sh <ACR_NAME> <AKS_RESOURCE_GROUP> <AKS_CLUSTER_NAME>
+
+set -e
+
+# 颜色输出
+RED='\033[0;31m'
+GREEN='\033[0;32m'
+YELLOW='\033[1;33m'
+NC='\033[0m' # No Color
+
+# 检查参数
+if [ $# -lt 3 ]; then
+    echo -e "${RED}错误: 缺少参数${NC}"
+    echo "使用方法: $0 <ACR_NAME> <AKS_RESOURCE_GROUP> <AKS_CLUSTER_NAME>"
+    echo "示例: $0 myregistry myresourcegroup myakscluster"
+    exit 1
+fi
+
+ACR_NAME=$1
+AKS_RESOURCE_GROUP=$2
+AKS_CLUSTER_NAME=$3
+
+echo -e "${GREEN}开始部署到 AKS...${NC}"
+echo "ACR 名称: $ACR_NAME"
+echo "资源组: $AKS_RESOURCE_GROUP"
+echo "AKS 集群: $AKS_CLUSTER_NAME"
+
+# 检查 kubectl 是否安装
+if ! command -v kubectl &> /dev/null; then
+    echo -e "${RED}错误: kubectl 未安装${NC}"
+    exit 1
+fi
+
+# 检查 az 是否安装
+if ! command -v az &> /dev/null; then
+    echo -e "${RED}错误: Azure CLI 未安装${NC}"
+    exit 1
+fi
+
+# 获取 AKS 凭据
+echo -e "${YELLOW}获取 AKS 集群凭据...${NC}"
+az aks get-credentials --resource-group $AKS_RESOURCE_GROUP --name $AKS_CLUSTER_NAME --overwrite-existing
+
+# 检查是否已连接到集群
+if ! kubectl cluster-info &> /dev/null; then
+    echo -e "${RED}错误: 无法连接到 AKS 集群${NC}"
+    exit 1
+fi
+
+# 创建命名空间
+echo -e "${YELLOW}创建命名空间...${NC}"
+kubectl apply -f namespace.yaml
+
+# 创建 ACR 拉取密钥（如果需要）
+echo -e "${YELLOW}配置 ACR 访问...${NC}"
+# 检查 AKS 是否已附加 ACR
+ACR_ATTACHED=$(az aks show -n $AKS_CLUSTER_NAME -g $AKS_RESOURCE_GROUP --query "servicePrincipalProfile" -o tsv)
+if [ -z "$ACR_ATTACHED" ]; then
+    echo "附加 ACR 到 AKS..."
+    az aks update -n $AKS_CLUSTER_NAME -g $AKS_RESOURCE_GROUP --attach-acr $ACR_NAME
+else
+    echo "ACR 已附加，跳过..."
+fi
+
+# 创建 Secret（提示用户）
+echo -e "${YELLOW}请确保已创建 Secret:${NC}"
+echo "  kubectl create secret generic taiji-secrets --from-literal=... --namespace=taiji-ai"
+echo "  或使用 Azure Key Vault Provider"
+read -p "是否已创建 Secret? (y/n) " -n 1 -r
+echo
+if [[ ! $REPLY =~ ^[Yy]$ ]]; then
+    echo -e "${RED}请先创建 Secret 后再继续${NC}"
+    exit 1
+fi
+
+# 替换配置文件中的变量
+echo -e "${YELLOW}替换配置变量...${NC}"
+export ACR_NAME=$ACR_NAME
+
+# 应用 ConfigMap
+echo -e "${YELLOW}应用 ConfigMap...${NC}"
+kubectl apply -f configmap.yaml
+
+# 应用所有部署配置
+echo -e "${YELLOW}部署 NATS...${NC}"
+envsubst < nats-deployment.yaml | kubectl apply -f -
+
+echo -e "${YELLOW}部署 Model Gateway...${NC}"
+envsubst < model-gateway-deployment.yaml | kubectl apply -f -
+
+echo -e "${YELLOW}部署 Data Ingestion...${NC}"
+envsubst < data-ingestion-deployment.yaml | kubectl apply -f -
+
+echo -e "${YELLOW}部署 MCP Server...${NC}"
+envsubst < mcp-server-deployment.yaml | kubectl apply -f -
+
+echo -e "${YELLOW}部署监控服务...${NC}"
+kubectl apply -f monitoring-deployment.yaml
+
+# 等待部署完成
+echo -e "${YELLOW}等待部署就绪...${NC}"
+kubectl wait --for=condition=available --timeout=300s deployment/nats -n taiji-ai || true
+kubectl wait --for=condition=available --timeout=300s deployment/litellm-gateway -n taiji-ai || true
+kubectl wait --for=condition=available --timeout=300s deployment/data-ingestion -n taiji-ai || true
+kubectl wait --for=condition=available --timeout=300s deployment/mcp-server -n taiji-ai || true
+
+# 应用 Ingress（可选）
+read -p "是否部署 Ingress? (y/n) " -n 1 -r
+echo
+if [[ $REPLY =~ ^[Yy]$ ]]; then
+    echo -e "${YELLOW}部署 Ingress...${NC}"
+    kubectl apply -f ingress.yaml
+fi
+
+# 显示部署状态
+echo -e "${GREEN}部署完成！${NC}"
+echo ""
+echo "查看部署状态:"
+echo "  kubectl get pods -n taiji-ai"
+echo "  kubectl get services -n taiji-ai"
+echo "  kubectl get ingress -n taiji-ai"
+echo ""
+echo "查看日志:"
+echo "  kubectl logs -f deployment/mcp-server -n taiji-ai"
+echo "  kubectl logs -f deployment/litellm-gateway -n taiji-ai"
+echo "  kubectl logs -f deployment/data-ingestion -n taiji-ai"
+
diff --git a/k8s/ingress.yaml b/k8s/ingress.yaml
new file mode 100644
index 0000000..6445d2d
--- /dev/null
+++ b/k8s/ingress.yaml
@@ -0,0 +1,147 @@
+# Ingress 配置 - 使用 Azure Application Gateway 或 NGINX Ingress
+# 根据您的 AKS 配置选择相应的 Ingress Controller
+
+---
+# 选项1: Azure Application Gateway Ingress Controller (AGIC)
+apiVersion: networking.k8s.io/v1
+kind: Ingress
+metadata:
+  name: taiji-ingress
+  namespace: taiji-ai
+  annotations:
+    # Azure Application Gateway Ingress Controller 注解
+    kubernetes.io/ingress.class: azure/application-gateway
+    appgw.ingress.kubernetes.io/ssl-redirect: "true"
+    appgw.ingress.kubernetes.io/connection-draining: "true"
+    appgw.ingress.kubernetes.io/connection-draining-timeout: "30"
+    appgw.ingress.kubernetes.io/backend-protocol: "http"
+    appgw.ingress.kubernetes.io/request-timeout: "300"
+    appgw.ingress.kubernetes.io/health-probe-path: "/health"
+    appgw.ingress.kubernetes.io/health-probe-interval: "30"
+    appgw.ingress.kubernetes.io/health-probe-timeout: "10"
+    appgw.ingress.kubernetes.io/health-probe-unhealthy-threshold: "3"
+    
+    # CORS 配置
+    appgw.ingress.kubernetes.io/cors-allow-origin: "*"
+    appgw.ingress.kubernetes.io/cors-allow-methods: "GET,POST,PUT,DELETE,OPTIONS"
+    appgw.ingress.kubernetes.io/cors-allow-headers: "*"
+    
+    # 证书配置（使用 Azure Key Vault）
+    # cert-manager.io/cluster-issuer: "letsencrypt-prod"
+spec:
+  tls:
+  - hosts:
+    - api.taiji-ai.com
+    - gateway.taiji-ai.com
+    secretName: taiji-tls-cert
+  rules:
+  # MCP Server API
+  - host: api.taiji-ai.com
+    http:
+      paths:
+      - path: /api
+        pathType: Prefix
+        backend:
+          service:
+            name: mcp-server
+            port:
+              number: 8000
+      - path: /health
+        pathType: Exact
+        backend:
+          service:
+            name: mcp-server
+            port:
+              number: 8000
+      - path: /docs
+        pathType: Prefix
+        backend:
+          service:
+            name: mcp-server
+            port:
+              number: 8000
+      - path: /openapi.json
+        pathType: Exact
+        backend:
+          service:
+            name: mcp-server
+            port:
+              number: 8000
+  
+  # Data Ingestion API
+  - host: api.taiji-ai.com
+    http:
+      paths:
+      - path: /ingestion
+        pathType: Prefix
+        backend:
+          service:
+            name: data-ingestion
+            port:
+              number: 8000
+  
+  # Model Gateway (内部使用，可选暴露)
+  - host: gateway.taiji-ai.com
+    http:
+      paths:
+      - path: /
+        pathType: Prefix
+        backend:
+          service:
+            name: litellm-gateway
+            port:
+              number: 4000
+
+---
+# 选项2: NGINX Ingress Controller (如果使用 NGINX 而不是 AGIC)
+# 取消注释以下配置，并注释掉上面的 AGIC 配置
+# apiVersion: networking.k8s.io/v1
+# kind: Ingress
+# metadata:
+#   name: taiji-ingress-nginx
+#   namespace: taiji-ai
+#   annotations:
+#     kubernetes.io/ingress.class: nginx
+#     cert-manager.io/cluster-issuer: letsencrypt-prod
+#     nginx.ingress.kubernetes.io/ssl-redirect: "true"
+#     nginx.ingress.kubernetes.io/proxy-body-size: "50m"
+#     nginx.ingress.kubernetes.io/proxy-read-timeout: "300"
+#     nginx.ingress.kubernetes.io/proxy-send-timeout: "300"
+#     nginx.ingress.kubernetes.io/cors-allow-origin: "*"
+#     nginx.ingress.kubernetes.io/cors-allow-methods: "GET,POST,PUT,DELETE,OPTIONS"
+#     nginx.ingress.kubernetes.io/cors-allow-headers: "*"
+# spec:
+#   tls:
+#   - hosts:
+#     - api.taiji-ai.com
+#     - gateway.taiji-ai.com
+#     secretName: taiji-tls-cert
+#   rules:
+#   - host: api.taiji-ai.com
+#     http:
+#       paths:
+#       - path: /api
+#         pathType: Prefix
+#         backend:
+#           service:
+#             name: mcp-server
+#             port:
+#               number: 8000
+#       - path: /ingestion
+#         pathType: Prefix
+#         backend:
+#           service:
+#             name: data-ingestion
+#             port:
+#               number: 8000
+#   - host: gateway.taiji-ai.com
+#     http:
+#       paths:
+#       - path: /
+#         pathType: Prefix
+#         backend:
+#           service:
+#             name: litellm-gateway
+#             port:
+#               number: 4000
+
diff --git a/k8s/mcp-server-deployment.yaml b/k8s/mcp-server-deployment.yaml
new file mode 100644
index 0000000..9ef1016
--- /dev/null
+++ b/k8s/mcp-server-deployment.yaml
@@ -0,0 +1,183 @@
+# MCP Server 部署配置
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: mcp-server
+  namespace: taiji-ai
+  labels:
+    app: mcp-server
+    component: server
+spec:
+  replicas: 3
+  selector:
+    matchLabels:
+      app: mcp-server
+  template:
+    metadata:
+      labels:
+        app: mcp-server
+        component: server
+    spec:
+      containers:
+      - name: mcp-server
+        image: ${ACR_NAME}.azurecr.io/mcp-server:latest
+        imagePullPolicy: Always
+        ports:
+        - containerPort: 8000
+          name: http
+          protocol: TCP
+        env:
+        - name: DATABASE_URL
+          valueFrom:
+            secretKeyRef:
+              name: taiji-secrets
+              key: database-url
+        - name: ASYNC_DATABASE_URL
+          valueFrom:
+            secretKeyRef:
+              name: taiji-secrets
+              key: async-database-url
+        - name: REDIS_URL
+          valueFrom:
+            secretKeyRef:
+              name: taiji-secrets
+              key: redis-url
+        - name: NATS_URL
+          valueFrom:
+            configMapKeyRef:
+              name: taiji-config
+              key: nats-url
+        - name: LITELLM_URL
+          valueFrom:
+            configMapKeyRef:
+              name: taiji-config
+              key: litellm-url
+        - name: LITELLM_API_KEY
+          valueFrom:
+            secretKeyRef:
+              name: taiji-secrets
+              key: litellm-api-key
+        - name: JWT_SECRET
+          valueFrom:
+            secretKeyRef:
+              name: taiji-secrets
+              key: jwt-secret
+        - name: ENCRYPTION_KEY
+          valueFrom:
+            secretKeyRef:
+              name: taiji-secrets
+              key: encryption-key
+        - name: AZURE_STORAGE_CONNECTION_STRING
+          valueFrom:
+            secretKeyRef:
+              name: taiji-secrets
+              key: azure-storage-connection-string
+              optional: true
+        - name: ENVIRONMENT
+          valueFrom:
+            configMapKeyRef:
+              name: taiji-config
+              key: environment
+        - name: DEBUG
+          valueFrom:
+            configMapKeyRef:
+              name: taiji-config
+              key: debug
+        - name: LOG_LEVEL
+          valueFrom:
+            configMapKeyRef:
+              name: taiji-config
+              key: log-level
+        - name: PYTHONUNBUFFERED
+          value: "1"
+        resources:
+          requests:
+            memory: "512Mi"
+            cpu: "250m"
+          limits:
+            memory: "1Gi"
+            cpu: "500m"
+        volumeMounts:
+        - name: logs
+          mountPath: /app/logs
+        livenessProbe:
+          httpGet:
+            path: /health
+            port: 8000
+          initialDelaySeconds: 30
+          periodSeconds: 10
+          timeoutSeconds: 5
+          failureThreshold: 3
+        readinessProbe:
+          httpGet:
+            path: /health
+            port: 8000
+          initialDelaySeconds: 10
+          periodSeconds: 5
+          timeoutSeconds: 3
+          failureThreshold: 2
+      volumes:
+      - name: logs
+        emptyDir: {}
+      imagePullSecrets:
+      - name: acr-secret
+
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: mcp-server
+  namespace: taiji-ai
+  labels:
+    app: mcp-server
+spec:
+  type: ClusterIP
+  ports:
+  - port: 8000
+    targetPort: 8000
+    protocol: TCP
+    name: http
+  selector:
+    app: mcp-server
+
+---
+# HorizontalPodAutoscaler
+apiVersion: autoscaling/v2
+kind: HorizontalPodAutoscaler
+metadata:
+  name: mcp-server-hpa
+  namespace: taiji-ai
+spec:
+  scaleTargetRef:
+    apiVersion: apps/v1
+    kind: Deployment
+    name: mcp-server
+  minReplicas: 3
+  maxReplicas: 10
+  metrics:
+  - type: Resource
+    resource:
+      name: cpu
+      target:
+        type: Utilization
+        averageUtilization: 70
+  - type: Resource
+    resource:
+      name: memory
+      target:
+        type: Utilization
+        averageUtilization: 80
+
+---
+# PodDisruptionBudget - 确保高可用
+apiVersion: policy/v1
+kind: PodDisruptionBudget
+metadata:
+  name: mcp-server-pdb
+  namespace: taiji-ai
+spec:
+  minAvailable: 2
+  selector:
+    matchLabels:
+      app: mcp-server
+
diff --git a/k8s/model-gateway-deployment.yaml b/k8s/model-gateway-deployment.yaml
new file mode 100644
index 0000000..594d753
--- /dev/null
+++ b/k8s/model-gateway-deployment.yaml
@@ -0,0 +1,188 @@
+# LiteLLM Model Gateway 部署配置
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: litellm-gateway
+  namespace: taiji-ai
+  labels:
+    app: litellm-gateway
+    component: gateway
+spec:
+  replicas: 2
+  selector:
+    matchLabels:
+      app: litellm-gateway
+  template:
+    metadata:
+      labels:
+        app: litellm-gateway
+        component: gateway
+    spec:
+      containers:
+      - name: litellm-gateway
+        image: ${ACR_NAME}.azurecr.io/litellm-gateway:latest
+        imagePullPolicy: Always
+        ports:
+        - containerPort: 4000
+          name: http
+          protocol: TCP
+        env:
+        - name: LITELLM_MASTER_KEY
+          valueFrom:
+            secretKeyRef:
+              name: taiji-secrets
+              key: litellm-master-key
+        - name: DATABASE_URL
+          valueFrom:
+            secretKeyRef:
+              name: taiji-secrets
+              key: database-url
+        - name: REDIS_URL
+          valueFrom:
+            secretKeyRef:
+              name: taiji-secrets
+              key: redis-url
+        - name: OPENROUTER_API_KEY
+          valueFrom:
+            secretKeyRef:
+              name: taiji-secrets
+              key: openrouter-api-key
+        - name: OPENROUTER_BASE_URL
+          valueFrom:
+            configMapKeyRef:
+              name: taiji-config
+              key: openrouter-base-url
+        - name: LITELLM_CONFIG_PATH
+          valueFrom:
+            configMapKeyRef:
+              name: taiji-config
+              key: litellm-config-path
+        resources:
+          requests:
+            memory: "512Mi"
+            cpu: "250m"
+          limits:
+            memory: "1Gi"
+            cpu: "1000m"
+        volumeMounts:
+        - name: litellm-config
+          mountPath: /app/config
+          readOnly: true
+        - name: logs
+          mountPath: /app/logs
+        livenessProbe:
+          httpGet:
+            path: /health
+            port: 4000
+          initialDelaySeconds: 60
+          periodSeconds: 30
+          timeoutSeconds: 25
+          failureThreshold: 3
+        readinessProbe:
+          httpGet:
+            path: /health
+            port: 4000
+          initialDelaySeconds: 30
+          periodSeconds: 10
+          timeoutSeconds: 5
+          failureThreshold: 2
+      volumes:
+      - name: litellm-config
+        configMap:
+          name: litellm-config
+      - name: logs
+        emptyDir: {}
+      imagePullSecrets:
+      - name: acr-secret
+
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: litellm-gateway
+  namespace: taiji-ai
+  labels:
+    app: litellm-gateway
+spec:
+  type: ClusterIP
+  ports:
+  - port: 4000
+    targetPort: 4000
+    protocol: TCP
+    name: http
+  selector:
+    app: litellm-gateway
+
+---
+# LiteLLM 配置文件 ConfigMap
+apiVersion: v1
+kind: ConfigMap
+metadata:
+  name: litellm-config
+  namespace: taiji-ai
+data:
+  litellm_simple.yaml: |
+    model_list:
+      - model_name: gpt-3.5-turbo
+        litellm_params:
+          model: openrouter/gpt-3.5-turbo
+          api_key: os.environ/OPENROUTER_API_KEY
+          api_base: https://openrouter.ai/api/v1
+      
+      - model_name: openrouter-gpt-3.5-turbo
+        litellm_params:
+          model: openrouter/gpt-3.5-turbo
+          api_key: os.environ/OPENROUTER_API_KEY
+          api_base: https://openrouter.ai/api/v1
+      
+      - model_name: openrouter-gpt-4o-mini
+        litellm_params:
+          model: openrouter/gpt-4o-mini
+          api_key: os.environ/OPENROUTER_API_KEY
+          api_base: https://openrouter.ai/api/v1
+      
+      - model_name: openrouter-claude-3.5-sonnet
+        litellm_params:
+          model: openrouter/anthropic/claude-3.5-sonnet
+          api_key: os.environ/OPENROUTER_API_KEY
+          api_base: https://openrouter.ai/api/v1
+      
+      - model_name: openrouter-claude-3-opus
+        litellm_params:
+          model: openrouter/anthropic/claude-3-opus
+          api_key: os.environ/OPENROUTER_API_KEY
+          api_base: https://openrouter.ai/api/v1
+    
+    general_settings:
+      master_key: os.environ/LITELLM_MASTER_KEY
+      database_url: os.environ/DATABASE_URL
+      redis_url: os.environ/REDIS_URL
+
+---
+# HorizontalPodAutoscaler
+apiVersion: autoscaling/v2
+kind: HorizontalPodAutoscaler
+metadata:
+  name: litellm-gateway-hpa
+  namespace: taiji-ai
+spec:
+  scaleTargetRef:
+    apiVersion: apps/v1
+    kind: Deployment
+    name: litellm-gateway
+  minReplicas: 2
+  maxReplicas: 10
+  metrics:
+  - type: Resource
+    resource:
+      name: cpu
+      target:
+        type: Utilization
+        averageUtilization: 70
+  - type: Resource
+    resource:
+      name: memory
+      target:
+        type: Utilization
+        averageUtilization: 80
+
diff --git a/k8s/monitoring-deployment.yaml b/k8s/monitoring-deployment.yaml
new file mode 100644
index 0000000..a6e0b9f
--- /dev/null
+++ b/k8s/monitoring-deployment.yaml
@@ -0,0 +1,269 @@
+# Prometheus 和 Grafana 监控服务部署配置
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: prometheus
+  namespace: taiji-ai
+  labels:
+    app: prometheus
+    component: monitoring
+spec:
+  replicas: 1
+  selector:
+    matchLabels:
+      app: prometheus
+  template:
+    metadata:
+      labels:
+        app: prometheus
+        component: monitoring
+    spec:
+      containers:
+      - name: prometheus
+        image: prom/prometheus:latest
+        ports:
+        - containerPort: 9090
+          name: http
+          protocol: TCP
+        args:
+        - '--config.file=/etc/prometheus/prometheus.yml'
+        - '--storage.tsdb.path=/prometheus'
+        - '--web.console.libraries=/etc/prometheus/console_libraries'
+        - '--web.console.templates=/etc/prometheus/consoles'
+        - '--web.enable-lifecycle'
+        resources:
+          requests:
+            memory: "512Mi"
+            cpu: "250m"
+          limits:
+            memory: "2Gi"
+            cpu: "1000m"
+        volumeMounts:
+        - name: prometheus-config
+          mountPath: /etc/prometheus
+          readOnly: true
+        - name: prometheus-data
+          mountPath: /prometheus
+      volumes:
+      - name: prometheus-config
+        configMap:
+          name: prometheus-config
+      - name: prometheus-data
+        persistentVolumeClaim:
+          claimName: prometheus-pvc
+
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: prometheus
+  namespace: taiji-ai
+  labels:
+    app: prometheus
+spec:
+  type: ClusterIP
+  ports:
+  - port: 9090
+    targetPort: 9090
+    protocol: TCP
+    name: http
+  selector:
+    app: prometheus
+
+---
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: grafana
+  namespace: taiji-ai
+  labels:
+    app: grafana
+    component: monitoring
+spec:
+  replicas: 1
+  selector:
+    matchLabels:
+      app: grafana
+  template:
+    metadata:
+      labels:
+        app: grafana
+        component: monitoring
+    spec:
+      containers:
+      - name: grafana
+        image: grafana/grafana:latest
+        ports:
+        - containerPort: 3000
+          name: http
+          protocol: TCP
+        env:
+        - name: GF_SECURITY_ADMIN_PASSWORD
+          valueFrom:
+            secretKeyRef:
+              name: taiji-secrets
+              key: grafana-admin-password
+              optional: true
+        - name: GF_SERVER_ROOT_URL
+          value: "http://grafana.taiji-ai.com"
+        resources:
+          requests:
+            memory: "256Mi"
+            cpu: "100m"
+          limits:
+            memory: "512Mi"
+            cpu: "500m"
+        volumeMounts:
+        - name: grafana-data
+          mountPath: /var/lib/grafana
+        - name: grafana-dashboards
+          mountPath: /etc/grafana/provisioning/dashboards
+        - name: grafana-datasources
+          mountPath: /etc/grafana/provisioning/datasources
+      volumes:
+      - name: grafana-data
+        persistentVolumeClaim:
+          claimName: grafana-pvc
+      - name: grafana-dashboards
+        configMap:
+          name: grafana-dashboards
+      - name: grafana-datasources
+        configMap:
+          name: grafana-datasources
+
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: grafana
+  namespace: taiji-ai
+  labels:
+    app: grafana
+spec:
+  type: ClusterIP
+  ports:
+  - port: 3000
+    targetPort: 3000
+    protocol: TCP
+    name: http
+  selector:
+    app: grafana
+
+---
+# Prometheus 配置
+apiVersion: v1
+kind: ConfigMap
+metadata:
+  name: prometheus-config
+  namespace: taiji-ai
+data:
+  prometheus.yml: |
+    global:
+      scrape_interval: 15s
+      evaluation_interval: 15s
+    
+    scrape_configs:
+      - job_name: 'prometheus'
+        static_configs:
+          - targets: ['localhost:9090']
+      
+      - job_name: 'litellm-gateway'
+        kubernetes_sd_configs:
+          - role: endpoints
+            namespaces:
+              names:
+                - taiji-ai
+        relabel_configs:
+          - source_labels: [__meta_kubernetes_service_name]
+            action: keep
+            regex: litellm-gateway
+      
+      - job_name: 'mcp-server'
+        kubernetes_sd_configs:
+          - role: endpoints
+            namespaces:
+              names:
+                - taiji-ai
+        relabel_configs:
+          - source_labels: [__meta_kubernetes_service_name]
+            action: keep
+            regex: mcp-server
+      
+      - job_name: 'data-ingestion'
+        kubernetes_sd_configs:
+          - role: endpoints
+            namespaces:
+              names:
+                - taiji-ai
+        relabel_configs:
+          - source_labels: [__meta_kubernetes_service_name]
+            action: keep
+            regex: data-ingestion
+
+---
+# Grafana 数据源配置
+apiVersion: v1
+kind: ConfigMap
+metadata:
+  name: grafana-datasources
+  namespace: taiji-ai
+data:
+  prometheus.yaml: |
+    apiVersion: 1
+    datasources:
+      - name: Prometheus
+        type: prometheus
+        access: proxy
+        url: http://prometheus:9090
+        isDefault: true
+        editable: true
+
+---
+# Grafana 仪表板配置
+apiVersion: v1
+kind: ConfigMap
+metadata:
+  name: grafana-dashboards
+  namespace: taiji-ai
+data:
+  dashboard-provider.yaml: |
+    apiVersion: 1
+    providers:
+      - name: 'Default'
+        orgId: 1
+        folder: ''
+        type: file
+        disableDeletion: false
+        editable: true
+        options:
+          path: /etc/grafana/provisioning/dashboards
+
+---
+# 持久化存储
+apiVersion: v1
+kind: PersistentVolumeClaim
+metadata:
+  name: prometheus-pvc
+  namespace: taiji-ai
+spec:
+  accessModes:
+    - ReadWriteOnce
+  storageClassName: managed-csi
+  resources:
+    requests:
+      storage: 50Gi
+
+---
+apiVersion: v1
+kind: PersistentVolumeClaim
+metadata:
+  name: grafana-pvc
+  namespace: taiji-ai
+spec:
+  accessModes:
+    - ReadWriteOnce
+  storageClassName: managed-csi
+  resources:
+    requests:
+      storage: 10Gi
+
diff --git a/k8s/namespace.yaml b/k8s/namespace.yaml
new file mode 100644
index 0000000..6625655
--- /dev/null
+++ b/k8s/namespace.yaml
@@ -0,0 +1,8 @@
+apiVersion: v1
+kind: Namespace
+metadata:
+  name: taiji-ai
+  labels:
+    name: taiji-ai
+    environment: production
+
diff --git a/k8s/nats-deployment.yaml b/k8s/nats-deployment.yaml
new file mode 100644
index 0000000..a9da60e
--- /dev/null
+++ b/k8s/nats-deployment.yaml
@@ -0,0 +1,111 @@
+# NATS 消息队列部署配置
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: nats
+  namespace: taiji-ai
+  labels:
+    app: nats
+    component: messaging
+spec:
+  replicas: 3
+  selector:
+    matchLabels:
+      app: nats
+  template:
+    metadata:
+      labels:
+        app: nats
+        component: messaging
+    spec:
+      containers:
+      - name: nats
+        image: nats:2.10-alpine
+        ports:
+        - containerPort: 4222
+          name: client
+          protocol: TCP
+        - containerPort: 6222
+          name: routing
+          protocol: TCP
+        - containerPort: 8222
+          name: monitoring
+          protocol: TCP
+        command:
+        - "/nats-server"
+        args:
+        - "-js"
+        - "-m"
+        - "8222"
+        - "-cluster"
+        - "nats://0.0.0.0:6222"
+        - "-routes"
+        - "nats://nats-0.nats.taiji-ai.svc.cluster.local:6222,nats://nats-1.nats.taiji-ai.svc.cluster.local:6222,nats://nats-2.nats.taiji-ai.svc.cluster.local:6222"
+        resources:
+          requests:
+            memory: "256Mi"
+            cpu: "100m"
+          limits:
+            memory: "512Mi"
+            cpu: "500m"
+        volumeMounts:
+        - name: nats-data
+          mountPath: /data
+        livenessProbe:
+          httpGet:
+            path: /healthz
+            port: 8222
+          initialDelaySeconds: 10
+          periodSeconds: 10
+        readinessProbe:
+          httpGet:
+            path: /healthz
+            port: 8222
+          initialDelaySeconds: 5
+          periodSeconds: 5
+      volumes:
+      - name: nats-data
+        persistentVolumeClaim:
+          claimName: nats-pvc
+
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: nats
+  namespace: taiji-ai
+  labels:
+    app: nats
+spec:
+  type: ClusterIP
+  ports:
+  - port: 4222
+    targetPort: 4222
+    protocol: TCP
+    name: client
+  - port: 6222
+    targetPort: 6222
+    protocol: TCP
+    name: routing
+  - port: 8222
+    targetPort: 8222
+    protocol: TCP
+    name: monitoring
+  selector:
+    app: nats
+
+---
+# NATS 持久化存储
+apiVersion: v1
+kind: PersistentVolumeClaim
+metadata:
+  name: nats-pvc
+  namespace: taiji-ai
+spec:
+  accessModes:
+    - ReadWriteOnce
+  storageClassName: managed-csi  # Azure 托管存储类
+  resources:
+    requests:
+      storage: 10Gi
+
diff --git a/k8s/secrets.yaml.template b/k8s/secrets.yaml.template
new file mode 100644
index 0000000..d13b4b8
--- /dev/null
+++ b/k8s/secrets.yaml.template
@@ -0,0 +1,70 @@
+# Kubernetes Secrets 模板
+# 注意：这是一个模板文件，实际部署时请使用以下方式创建 Secret：
+#
+# 方式1: 使用 kubectl 命令
+# kubectl create secret generic taiji-secrets \
+#   --from-literal=database-url="postgresql+asyncpg://..." \
+#   --from-literal=redis-url="rediss://..." \
+#   --namespace=taiji-ai
+#
+# 方式2: 使用 Azure Key Vault (推荐)
+# 安装 Azure Key Vault Provider: https://azure.github.io/secrets-store-csi-driver-provider-azure/
+#
+# 方式3: 使用 sealed-secrets (推荐用于 GitOps)
+# https://github.com/bitnami-labs/sealed-secrets
+
+apiVersion: v1
+kind: Secret
+metadata:
+  name: taiji-secrets
+  namespace: taiji-ai
+type: Opaque
+stringData:
+  # 数据库连接字符串 (Azure Database for PostgreSQL)
+  database-url: "postgresql+asyncpg://USER:PASSWORD@SERVER.postgres.database.azure.com:5432/DATABASE?sslmode=require"
+  async-database-url: "postgresql+asyncpg://USER:PASSWORD@SERVER.postgres.database.azure.com:5432/DATABASE?sslmode=require"
+  
+  # Redis连接字符串 (Azure Cache for Redis)
+  redis-url: "rediss://:PASSWORD@SERVER.redis.cache.windows.net:6380/0?ssl_cert_reqs=required"
+  
+  # JWT密钥
+  jwt-secret: "CHANGE_THIS_SECRET_KEY_IN_PRODUCTION_USE_STRONG_RANDOM_STRING"
+  
+  # 加密密钥
+  encryption-key: "CHANGE_THIS_ENCRYPTION_KEY_USE_STRONG_RANDOM_STRING"
+  
+  # LiteLLM API密钥
+  litellm-master-key: "sk-taiji-master-key"
+  litellm-api-key: "sk-taiji-master-key"
+  
+  # OpenRouter API密钥
+  openrouter-api-key: "YOUR_OPENROUTER_API_KEY"
+  
+  # RapidAPI配置
+  rapidapi-key: "YOUR_RAPIDAPI_KEY"
+  
+  # Azure存储连接字符串 (可选)
+  azure-storage-connection-string: "DefaultEndpointsProtocol=https;AccountName=ACCOUNT;AccountKey=KEY;EndpointSuffix=core.windows.net"
+
+---
+# ACR (Azure Container Registry) 拉取密钥
+# 创建方式：
+# kubectl create secret docker-registry acr-secret \
+#   --docker-server=${ACR_NAME}.azurecr.io \
+#   --docker-username=${SP_APP_ID} \
+#   --docker-password=${SP_PASSWORD} \
+#   --namespace=taiji-ai
+#
+# 或者使用 AKS 与 ACR 集成（推荐）：
+# az aks update -n <AKS_CLUSTER_NAME> -g <RESOURCE_GROUP> --attach-acr <ACR_NAME>
+
+apiVersion: v1
+kind: Secret
+metadata:
+  name: acr-secret
+  namespace: taiji-ai
+type: kubernetes.io/dockerconfigjson
+data:
+  .dockerconfigjson: <BASE64_ENCODED_DOCKER_CONFIG>
+  # 注意：实际使用时请使用上述命令创建，不要直接使用 base64 编码的配置
+
diff --git a/k8s/test-deployment.sh b/k8s/test-deployment.sh
new file mode 100755
index 0000000..c499061
--- /dev/null
+++ b/k8s/test-deployment.sh
@@ -0,0 +1,93 @@
+#!/bin/bash
+
+# 测试部署的服务
+# 使用方法: ./test-deployment.sh
+
+set -e
+
+export PATH=$HOME/bin:$PATH
+
+NAMESPACE="taiji-ai"
+
+echo "=========================================="
+echo "测试 AKS 部署的服务"
+echo "=========================================="
+echo ""
+
+# 1. 检查 Pod 状态
+echo "1. 检查 Pod 状态..."
+kubectl get pods -n $NAMESPACE
+echo ""
+
+# 2. 检查服务
+echo "2. 检查服务..."
+kubectl get services -n $NAMESPACE
+echo ""
+
+# 3. 测试 MCP Server 健康检查
+echo "3. 测试 MCP Server 健康检查..."
+MCP_POD=$(kubectl get pods -l app=mcp-server -n $NAMESPACE -o jsonpath='{.items[0].metadata.name}' 2>/dev/null || echo "")
+if [ -n "$MCP_POD" ]; then
+    echo "MCP Server Pod: $MCP_POD"
+    kubectl exec -n $NAMESPACE $MCP_POD -- curl -s http://localhost:8000/health || echo "健康检查失败"
+else
+    echo "未找到 MCP Server Pod"
+fi
+echo ""
+
+# 4. 测试 Data Ingestion 健康检查
+echo "4. 测试 Data Ingestion 健康检查..."
+DI_POD=$(kubectl get pods -l app=data-ingestion -n $NAMESPACE -o jsonpath='{.items[0].metadata.name}' 2>/dev/null || echo "")
+if [ -n "$DI_POD" ]; then
+    echo "Data Ingestion Pod: $DI_POD"
+    kubectl exec -n $NAMESPACE $DI_POD -- curl -s http://localhost:8000/health || echo "健康检查失败"
+else
+    echo "未找到 Data Ingestion Pod"
+fi
+echo ""
+
+# 5. 测试 LiteLLM Gateway 健康检查
+echo "5. 测试 LiteLLM Gateway 健康检查..."
+LG_POD=$(kubectl get pods -l app=litellm-gateway -n $NAMESPACE -o jsonpath='{.items[0].metadata.name}' 2>/dev/null || echo "")
+if [ -n "$LG_POD" ]; then
+    echo "LiteLLM Gateway Pod: $LG_POD"
+    kubectl exec -n $NAMESPACE $LG_POD -- curl -s -H "Authorization: Bearer sk-taiji-master-key" http://localhost:4000/health || echo "健康检查失败"
+else
+    echo "未找到 LiteLLM Gateway Pod"
+fi
+echo ""
+
+# 6. 测试服务端点（通过 port-forward）
+echo "6. 测试服务端点..."
+echo "注意: 以下测试需要手动运行 port-forward 命令"
+echo ""
+echo "测试 MCP Server:"
+echo "  kubectl port-forward svc/mcp-server 8002:8000 -n $NAMESPACE"
+echo "  curl http://localhost:8002/health"
+echo ""
+echo "测试 Data Ingestion:"
+echo "  kubectl port-forward svc/data-ingestion 8001:8000 -n $NAMESPACE"
+echo "  curl http://localhost:8001/health"
+echo ""
+echo "测试 LiteLLM Gateway:"
+echo "  kubectl port-forward svc/litellm-gateway 4000:4000 -n $NAMESPACE"
+echo "  curl -H 'Authorization: Bearer sk-taiji-master-key' http://localhost:4000/health"
+echo ""
+
+# 7. 检查日志（最近10行）
+echo "7. 检查服务日志（最近10行）..."
+echo ""
+echo "MCP Server 日志:"
+kubectl logs -l app=mcp-server -n $NAMESPACE --tail=10 2>&1 | head -10 || echo "无法获取日志"
+echo ""
+echo "Data Ingestion 日志:"
+kubectl logs -l app=data-ingestion -n $NAMESPACE --tail=10 2>&1 | head -10 || echo "无法获取日志"
+echo ""
+echo "LiteLLM Gateway 日志:"
+kubectl logs -l app=litellm-gateway -n $NAMESPACE --tail=10 2>&1 | head -10 || echo "无法获取日志"
+echo ""
+
+echo "=========================================="
+echo "测试完成"
+echo "=========================================="
+