部署 on 云原生 MCP 服务器

Kubernetes 部署

Mon, 01 Jan 0001 00:00:00 +0000

Kubernetes 部署#

本指南描述如何在 Kubernetes 集群中部署 Cloud Native MCP Server。

前提条件#

已配置的 Kubernetes 集群
kubectl 已安装并配置
适当的集群权限

基本部署#

Deployment#

创建部署清单：

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55


apiVersion: apps/v1
kind: Deployment
metadata:
 name: cloud-native-mcp-server
 namespace: default
spec:
 replicas: 1
 selector:
 matchLabels:
 app: cloud-native-mcp-server
 template:
 metadata:
 labels:
 app: cloud-native-mcp-server
 spec:
 serviceAccountName: cloud-native-mcp-server
 containers:
 - name: cloud-native-mcp-server
 image: mahmutabi/cloud-native-mcp-server:latest
 ports:
 - containerPort: 8080
 env:
 - name: MCP_MODE
 value: "sse"
 - name: MCP_ADDR
 value: "0.0.0.0:8080"
 - name: MCP_LOG_LEVEL
 value: "info"
 volumeMounts:
 - name: kubeconfig
 mountPath: /root/.kube
 readOnly: true
 resources:
 requests:
 memory: "256Mi"
 cpu: "250m"
 limits:
 memory: "512Mi"
 cpu: "500m"
 livenessProbe:
 httpGet:
 path: /health
 port: 8080
 initialDelaySeconds: 30
 periodSeconds: 10
 readinessProbe:
 httpGet:
 path: /health
 port: 8080
 initialDelaySeconds: 5
 periodSeconds: 5
 volumes:
 - name: kubeconfig
 configMap:
 name: kubeconfig

ServiceAccount 和 RBAC#

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27


apiVersion: v1
kind: ServiceAccount
metadata:
 name: cloud-native-mcp-server
 namespace: default
---
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRole
metadata:
 name: cloud-native-mcp-server
rules:
- apiGroups: ["*"]
 resources: ["*"]
 verbs: ["get", "list", "watch", "describe"]
---
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
 name: cloud-native-mcp-server
roleRef:
 apiGroup: rbac.authorization.k8s.io
 kind: ClusterRole
 name: cloud-native-mcp-server
subjects:
- kind: ServiceAccount
 name: cloud-native-mcp-server
 namespace: default

Service#

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13


apiVersion: v1
kind: Service
metadata:
 name: cloud-native-mcp-server
 namespace: default
spec:
 type: ClusterIP
 ports:
 - port: 8080
 targetPort: 8080
 protocol: TCP
 selector:
 app: cloud-native-mcp-server

Ingress#

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19


apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
 name: cloud-native-mcp-server
 namespace: default
 annotations:
 nginx.ingress.kubernetes.io/rewrite-target: /
spec:
 rules:
 - host: k8s-mcp.example.com
 http:
 paths:
 - path: /
 pathType: Prefix
 backend:
 service:
 name: cloud-native-mcp-server
 port:
 number: 8080

部署步骤#

 1
 2
 3
 4
 5
 6
 7
 8
 9
10


# 应用所有清单
kubectl apply -f deploy/kubernetes/

# 验证部署
kubectl get pods -l app=cloud-native-mcp-server
kubectl logs -l app=cloud-native-mcp-server

# 测试连接
kubectl port-forward svc/cloud-native-mcp-server 8080:8080
curl http://localhost:8080/health

高可用配置#

多副本部署#

1
2
3
4
5
6
7


apiVersion: apps/v1
kind: Deployment
metadata:
 name: cloud-native-mcp-server
spec:
 replicas: 3
 # ... 其他配置

自动扩缩容#

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24


apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
 name: cloud-native-mcp-server
spec:
 scaleTargetRef:
 apiVersion: apps/v1
 kind: Deployment
 name: cloud-native-mcp-server
 minReplicas: 3
 maxReplicas: 10
 metrics:
 - type: Resource
 resource:
 name: cpu
 target:
 type: Utilization
 averageUtilization: 80
 - type: Resource
 resource:
 name: memory
 target:
 type: Utilization
 averageUtilization: 80

配置管理#

使用 ConfigMap#

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17


apiVersion: v1
kind: ConfigMap
metadata:
 name: mcp-config
data:
 config.yaml: |
 server:
 mode: "sse"
 addr: "0.0.0.0:8080"
 logging:
 level: "info"
 kubernetes:
 kubeconfig: ""
 grafana:
 enabled: true
 url: "http://grafana:3000"
 apiKey: "${GRAFANA_API_KEY}"

使用 Secret#

1
2
3


kubectl create secret generic mcp-secrets \
 --from-literal=mcp-api-key='your-key' \
 --from-literal=grafana-api-key='your-grafana-key'

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11


env:
- name: MCP_AUTH_API_KEY
 valueFrom:
 secretKeyRef:
 name: mcp-secrets
 key: mcp-api-key
- name: GRAFANA_API_KEY
 valueFrom:
 secretKeyRef:
 name: mcp-secrets
 key: grafana-api-key

安全配置#

Pod 安全#

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13


spec:
 securityContext:
 runAsNonRoot: true
 runAsUser: 1000
 fsGroup: 1000
 containers:
 - name: cloud-native-mcp-server
 securityContext:
 allowPrivilegeEscalation: false
 capabilities:
 drop:
 - ALL
 readOnlyRootFilesystem: true

网络策略#

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23


apiVersion: networking.k8s.io/v1
kind: NetworkPolicy
metadata:
 name: cloud-native-mcp-server
spec:
 podSelector:
 matchLabels:
 app: cloud-native-mcp-server
 policyTypes:
 - Ingress
 - Egress
 ingress:
 - from:
 - namespaceSelector: {}
 ports:
 - protocol: TCP
 port: 8080
 egress:
 - to:
 - namespaceSelector: {}
 ports:
 - protocol: TCP
 port: 443

监控配置#

Service 监控#

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11


apiVersion: v1
kind: Service
metadata:
 name: cloud-native-mcp-server
 namespace: default
 annotations:
 prometheus.io/scrape: "true"
 prometheus.io/port: "8080"
 prometheus.io/path: "/metrics"
spec:
 # ... 其他配置

故障排查#

检查 Pod 状态#

1
2
3


kubectl get pods -l app=cloud-native-mcp-server
kubectl describe pod <pod-name>
kubectl logs <pod-name>

检查服务#

1
2


kubectl get svc cloud-native-mcp-server
kubectl describe svc cloud-native-mcp-server

检查 RBAC#

1
2
3
4


kubectl get sa cloud-native-mcp-server
kubectl get clusterrole cloud-native-mcp-server
kubectl get clusterrolebinding cloud-native-mcp-server
kubectl auth can-i list pods --as=system:serviceaccount:default:cloud-native-mcp-server

Docker 部署

Mon, 01 Jan 0001 00:00:00 +0000

Docker 部署#

本指南描述如何使用 Docker 部署 Cloud Native MCP Server。

前提条件#

Docker 20.10+ 已安装
适当的 Docker 权限

快速启动#

基本运行#

1
2
3
4
5


docker run -d \
 --name cloud-native-mcp-server \
 -p 8080:8080 \
 -v ~/.kube:/root/.kube:ro \
 mahmutabi/cloud-native-mcp-server:latest

自定义配置#

1
2
3
4
5
6
7
8
9


docker run -d \
 --name cloud-native-mcp-server \
 -p 8080:8080 \
 -v ~/.kube:/root/.kube:ro \
 -v $(pwd)/config.yaml:/app/config.yaml:ro \
 -e MCP_MODE=sse \
 -e MCP_ADDR=0.0.0.0:8080 \
 -e MCP_LOG_LEVEL=info \
 mahmutabi/cloud-native-mcp-server:latest

Docker Compose#

创建 docker-compose.yml#

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28


version: '3.8'

services:
 cloud-native-mcp-server:
 image: mahmutabi/cloud-native-mcp-server:latest
 container_name: cloud-native-mcp-server
 ports:
 - "8080:8080"
 volumes:
 - ~/.kube:/root/.kube:ro
 - ./config.yaml:/app/config.yaml:ro
 environment:
 - MCP_MODE=sse
 - MCP_ADDR=0.0.0.0:8080
 - MCP_LOG_LEVEL=info
 restart: unless-stopped
 healthcheck:
 test: ["CMD", "wget", "--spider", "http://localhost:8080/health"]
 interval: 30s
 timeout: 10s
 retries: 3
 start_period: 40s
 networks:
 - monitoring

networks:
 monitoring:
 external: true

使用 Docker Compose#

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11


# 启动
docker-compose up -d

# 查看日志
docker-compose logs -f

# 停止
docker-compose down

# 重启
docker-compose restart

配置文件#

config.yaml#

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24


server:
 mode: "sse"
 addr: "0.0.0.0:8080"

logging:
 level: "info"
 json: false

kubernetes:
 kubeconfig: ""

grafana:
 enabled: true
 url: "http://grafana:3000"
 apiKey: "${GRAFANA_API_KEY}"

prometheus:
 enabled: true
 address: "http://prometheus:9090"

auth:
 enabled: true
 mode: "apikey"
 apiKey: "${MCP_AUTH_API_KEY}"

环境变量#

支持的环境变量#

1
2
3
4
5
6


MCP_MODE=sse
MCP_ADDR=0.0.0.0:8080
MCP_LOG_LEVEL=info
MCP_AUTH_ENABLED=true
MCP_AUTH_MODE=apikey
MCP_AUTH_API_KEY=your-key

使用 .env 文件#

创建 .env 文件：

Helm 部署

Mon, 01 Jan 0001 00:00:00 +0000

Helm 部署#

本指南描述如何使用 Helm Chart 部署 Cloud Native MCP Server。

前提条件#

Helm 3.0+ 已安装
Kubernetes 集群已配置
kubectl 已安装并配置

添加仓库#

1
2
3
4
5
6
7
8


# 添加仓库
helm repo add k8s-mcp https://mahmut-Abi.github.io/cloud-native-mcp-server

# 更新仓库
helm repo update

# 搜索 Chart
helm search repo k8s-mcp

基本安装#

默认安装#

1
2
3
4
5
6
7
8


# 安装
helm install cloud-native-mcp-server k8s-mcp/cloud-native-mcp-server

# 升级
helm upgrade cloud-native-mcp-server k8s-mcp/cloud-native-mcp-server

# 卸载
helm uninstall cloud-native-mcp-server

指定命名空间#

1
2
3
4


# 安装到指定命名空间
helm install cloud-native-mcp-server k8s-mcp/cloud-native-mcp-server \
 --namespace monitoring \
 --create-namespace

自定义 Values#

创建 values.yaml#

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77


replicaCount: 2

image:
 repository: mahmutabi/cloud-native-mcp-server
 tag: latest
 pullPolicy: IfNotPresent

service:
 type: ClusterIP
 port: 8080

ingress:
 enabled: true
 annotations:
 cert-manager.io/cluster-issuer: "letsencrypt-prod"
 hosts:
 - host: k8s-mcp.example.com
 paths:
 - path: /
 pathType: Prefix
 tls:
 - secretName: k8s-mcp-tls
 hosts:
 - k8s-mcp.example.com

resources:
 requests:
 memory: "256Mi"
 cpu: "250m"
 limits:
 memory: "512Mi"
 cpu: "500m"

autoscaling:
 enabled: true
 minReplicas: 2
 maxReplicas: 10
 targetCPUUtilizationPercentage: 80
 targetMemoryUtilizationPercentage: 80

config:
 server:
 mode: "sse"
 addr: "0.0.0.0:8080"
 logging:
 level: "info"
 format: "json"
 kubernetes:
 kubeconfig: ""
 grafana:
 enabled: true
 url: "http://grafana:3000"
 apiKey: "${GRAFANA_API_KEY}"
 prometheus:
 enabled: true
 address: "http://prometheus:9090"
 auth:
 enabled: true
 mode: "apikey"
 apiKey: "${MCP_AUTH_API_KEY}"

rbac:
 create: true
 rules:
 - apiGroups: ["*"]
 resources: ["*"]
 verbs: ["get", "list", "watch", "describe"]

serviceAccount:
 create: true
 name: ""

nodeSelector: {}

tolerations: []

affinity: {}

使用自定义 Values 安装#

1
2
3
4


helm install cloud-native-mcp-server k8s-mcp/cloud-native-mcp-server \
 -f values.yaml \
 --namespace monitoring \
 --create-namespace

配置选项#

镜像配置#

1
2
3
4
5


image:
 repository: mahmutabi/cloud-native-mcp-server
 tag: "v1.0.0"
 pullPolicy: IfNotPresent
 pullSecrets: []

副本配置#

1
2
3
4
5
6
7
8


replicaCount: 3

autoscaling:
 enabled: true
 minReplicas: 3
 maxReplicas: 10
 targetCPUUtilizationPercentage: 80
 targetMemoryUtilizationPercentage: 80

资源配置#

1
2
3
4
5
6
7


resources:
 requests:
 memory: "512Mi"
 cpu: "500m"
 limits:
 memory: "1Gi"
 cpu: "1000m"

服务配置#

1
2
3
4
5
6
7


service:
 type: ClusterIP
 port: 8080
 annotations:
 prometheus.io/scrape: "true"
 prometheus.io/port: "8080"
 prometheus.io/path: "/metrics"

Ingress 配置#

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15


ingress:
 enabled: true
 className: "nginx"
 annotations:
 cert-manager.io/cluster-issuer: "letsencrypt-prod"
 nginx.ingress.kubernetes.io/ssl-redirect: "true"
 hosts:
 - host: k8s-mcp.example.com
 paths:
 - path: /
 pathType: Prefix
 tls:
 - secretName: k8s-mcp-tls
 hosts:
 - k8s-mcp.example.com

配置文件#

 1
 2
 3
 4
 5
 6
 7
 8
 9
10


config:
 server:
 mode: "sse"
 addr: "0.0.0.0:8080"
 logging:
 level: "info"
 format: "json"
 kubernetes:
 kubeconfig: ""
 # ... 其他服务配置

Secrets#

1
2
3
4
5
6


secrets:
 create: true
 mcpApiKey: "your-api-key"
 grafanaApiKey: "your-grafana-key"
 # 或者使用现有 Secret
 existingSecret: "mcp-secrets"

RBAC#

1
2
3
4
5
6
7
8
9


rbac:
 create: true
 rules:
 - apiGroups: [""]
 resources: ["pods", "services", "configmaps", "secrets"]
 verbs: ["get", "list", "watch"]
 - apiGroups: ["apps"]
 resources: ["deployments", "replicasets"]
 verbs: ["get", "list", "watch"]

ServiceAccount#

1
2
3
4


serviceAccount:
 create: true
 name: "cloud-native-mcp-server"
 annotations: {}

节点选择#

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19


nodeSelector:
 node.kubernetes.io/instance-type: "m5.large"

tolerations:
 - key: "workload"
 operator: "Equal"
 value: "monitoring"
 effect: "NoSchedule"

affinity:
 podAntiAffinity:
 requiredDuringSchedulingIgnoredDuringExecution:
 - labelSelector:
 matchExpressions:
 - key: app
 operator: In
 values:
 - cloud-native-mcp-server
 topologyKey: "kubernetes.io/hostname"

高级用法#

使用现有 ConfigMap#

1
2
3


configMap:
 create: false
 name: "mcp-config"

使用现有 Secret#

1
2
3


secret:
 create: false
 name: "mcp-secrets"

环境变量#

1
2
3
4
5
6
7
8


env:
 - name: CUSTOM_VAR
 value: "custom-value"
 - name: SECRET_VAR
 valueFrom:
 secretKeyRef:
 name: mcp-secrets
 key: secret-var

卷挂载#

 1
 2
 3
 4
 5
 6
 7
 8
 9
10


volumes:
 - name: kubeconfig
 configMap:
 name: kubeconfig
 optional: true

volumeMounts:
 - name: kubeconfig
 mountPath: /root/.kube
 readOnly: true

监控配置#

PodMonitor#

1
2
3
4
5


podMonitor:
 enabled: true
 interval: 30s
 scrapeTimeout: 10s
 namespace: monitoring

ServiceMonitor#

1
2
3
4
5


serviceMonitor:
 enabled: true
 interval: 30s
 scrapeTimeout: 10s
 namespace: monitoring

故障排查#

检查安装状态#

 1
 2
 3
 4
 5
 6
 7
 8
 9
10


helm status cloud-native-mcp-server -n monitoring

# 查看历史版本
helm history cloud-native-mcp-server -n monitoring

# 获取 Values
helm get values cloud-native-mcp-server -n monitoring

# 获取所有信息
helm get all cloud-native-mcp-server -n monitoring

调试安装#

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11


# Dry run
helm install cloud-native-mcp-server k8s-mcp/cloud-native-mcp-server \
 -f values.yaml \
 --namespace monitoring \
 --dry-run \
 --debug

# 渲染模板
helm template cloud-native-mcp-server k8s-mcp/cloud-native-mcp-server \
 -f values.yaml \
 --namespace monitoring

回滚#

1
2
3
4
5
6
7
8


# 查看历史
helm history cloud-native-mcp-server -n monitoring

# 回滚到上一个版本
helm rollback cloud-native-mcp-server -n monitoring

# 回滚到特定版本
helm rollback cloud-native-mcp-server 2 -n monitoring

部署 on 云原生 MCP 服务器

Kubernetes 部署

Kubernetes 部署#

前提条件#

基本部署#

Deployment#

ServiceAccount 和 RBAC#

Service#

Ingress#

部署步骤#

高可用配置#

多副本部署#

自动扩缩容#

配置管理#

使用 ConfigMap#

使用 Secret#

安全配置#

Pod 安全#

网络策略#

监控配置#

Service 监控#

故障排查#

检查 Pod 状态#

检查服务#

检查 RBAC#

相关文档#

Docker 部署

Docker 部署#

前提条件#

快速启动#

基本运行#

自定义配置#

Docker Compose#

创建 docker-compose.yml#

使用 Docker Compose#

配置文件#

config.yaml#

环境变量#

支持的环境变量#

使用 .env 文件#

Helm 部署

Helm 部署#

前提条件#

添加仓库#

基本安装#

默认安装#

指定命名空间#

自定义 Values#

创建 values.yaml#

使用自定义 Values 安装#

配置选项#

镜像配置#

副本配置#

资源配置#

服务配置#

Ingress 配置#

配置文件#

Secrets#

RBAC#

ServiceAccount#

节点选择#

高级用法#

使用现有 ConfigMap#

使用现有 Secret#

环境变量#

卷挂载#

监控配置#

PodMonitor#

ServiceMonitor#

故障排查#

检查安装状态#

调试安装#

回滚#

相关文档#