前言

Kubernetes集群對(duì)外提供服務(wù)時(shí)，Ingress是標(biāo)準(zhǔn)的服務(wù)暴露方式。Ingress資源定義了HTTP/HTTPS路由規(guī)則，而Ingress Controller則是這些規(guī)則的實(shí)現(xiàn)者。Ingress Controller負(fù)責(zé)監(jiān)控Ingress資源的變化，將路由規(guī)則翻譯為具體的反向代理配置，并動(dòng)態(tài)更新。

主流的Ingress Controller包括Nginx Ingress Controller、Traefik和Envoy Proxy。它們各有特點(diǎn)和適用場(chǎng)景，選擇合適的Ingress Controller對(duì)集群的穩(wěn)定性、性能和可維護(hù)性都有重要影響。

本文面向初中級(jí)運(yùn)維工程師，從架構(gòu)原理、配置方式、功能特性、性能表現(xiàn)、選型建議等維度對(duì)這三款I(lǐng)ngress Controller進(jìn)行深入對(duì)比。通過(guò)本文，讀者能夠理解每種Ingress Controller的工作機(jī)制，并根據(jù)自己的場(chǎng)景做出合理選擇。

第一章 Ingress基礎(chǔ)概念

1.1 Kubernetes網(wǎng)絡(luò)模型

Kubernetes采用扁平化的網(wǎng)絡(luò)模型，每個(gè)Pod都有獨(dú)立的IP地址，Pod之間可以直接通信，不需要NAT。Service是Pod的抽象，通過(guò)標(biāo)簽選擇器將一組Pod統(tǒng)一暴露，提供負(fù)載均衡和發(fā)現(xiàn)機(jī)制。

ClusterIP是Service的默認(rèn)類(lèi)型，僅在集群內(nèi)部可訪(fǎng)問(wèn)。要從外部訪(fǎng)問(wèn)服務(wù)，通常需要以下幾種方式：NodePort在每個(gè)節(jié)點(diǎn)上開(kāi)放一個(gè)端口；LoadBalancer需要云服務(wù)商支持，創(chuàng)建外部負(fù)載均衡器；Ingress則基于HTTP/HTTPS主機(jī)名和路徑提供七層路由。

1.2 Ingress資源定義

Ingress是Kubernetes的標(biāo)準(zhǔn)資源，用于配置HTTP/HTTPS外部訪(fǎng)問(wèn)路由。

apiVersion:networking.k8s.io/v1
kind:Ingress
metadata:
name:demo-ingress
annotations:
 nginx.ingress.kubernetes.io/rewrite-target:/
spec:
ingressClassName:nginx
rules:
 -host:demo.example.com
  http:
   paths:
    -path:/api
     pathType:Prefix
     backend:
      service:
       name:api-service
       port:
        number:80
    -path:/web
     pathType:Prefix
     backend:
       service:
       name:web-service
       port:
        number:8080
tls:
 -hosts:
   -demo.example.com
  secretName:demo-tls-secret

Ingress的配置通過(guò)注解（annotations）擴(kuò)展，不同的Ingress Controller支持不同的注解。ingressClassName字段指定使用的Ingress Controller類(lèi)，Kubernetes 1.18起支持此字段。

1.3 Ingress Controller職責(zé)

Ingress Controller的核心職責(zé)包括：監(jiān)控Kubernetes API獲取Ingress、Service、Endpoints等資源變化；將路由規(guī)則翻譯為反向代理配置；熱更新配置而不丟失請(qǐng)求；提供健康檢查和指標(biāo)暴露；處理TLS終止；實(shí)現(xiàn)負(fù)載均衡算法。

理解Ingress Controller的實(shí)現(xiàn)機(jī)制，有助于選擇合適的方案和排查問(wèn)題。每種Ingress Controller本質(zhì)上都是一個(gè)運(yùn)行在集群中的Deployment，它持續(xù)監(jiān)聽(tīng)Kubernetes API并更新自己的配置。

第二章 Nginx Ingress Controller

2.1 架構(gòu)與原理

Nginx Ingress Controller基于原生Nginx構(gòu)建，部署模式分為兩種：Deployment模式和DaemonSet模式。Deployment模式適合大規(guī)模集群，可以通過(guò)HPA實(shí)現(xiàn)自動(dòng)擴(kuò)縮容；DaemonSet模式在每個(gè)節(jié)點(diǎn)都運(yùn)行一個(gè)實(shí)例，適合低延遲要求的場(chǎng)景。

apiVersion:apps/v1
kind:Deployment
metadata:
name:nginx-ingress-controller
namespace:ingress-nginx
spec:
replicas:3
selector:
 matchLabels:
  app:nginx-ingress
template:
 metadata:
  labels:
   app:nginx-ingress
 spec:
  containers:
   -name:controller
    image:registry.k8s.io/ingress-nginx/controller:v1.9.4
    args:
     -/nginx-ingress-controller
     ---publish-service=$(POD_NAMESPACE)/ingress-nginx-controller
     ---election-id=ingress-controller-leader
     ---controller-class=k8s.io/ingress-nginx
     ---ingress-class=nginx
     ---configmap=$(POD_NAMESPACE)/nginx-configuration
    env:
     -name:POD_NAME
      valueFrom:
       fieldRef:
        fieldPath:metadata.name
     -name:POD_NAMESPACE
      valueFrom:
       fieldRef:
        fieldPath:metadata.namespace
    ports:
     -name:http
      containerPort:80
     -name:https
      containerPort:443
    livenessProbe:
     httpGet:
      path:/healthz
      port:10254
      scheme:HTTP
     initialDelaySeconds:10
     periodSeconds:10
    readinessProbe:
     httpGet:
      path:/healthz
      port:10254
      scheme:HTTP
     periodSeconds:5
    resources:
     requests:
      cpu:100m
      memory:90Mi
     limits:
      cpu:1
      memory:1Gi

Nginx Ingress Controller的工作流程：用戶(hù)創(chuàng)建或更新Ingress資源；Controller監(jiān)聽(tīng)到變化并生成Nginx配置文件；Nginx reload加載新配置；請(qǐng)求通過(guò)Service NodePort或LoadBalancer進(jìn)入Nginx。

2.2 配置方式

Nginx Ingress Controller支持兩種配置方式：通過(guò)Ingress注解和通過(guò)ConfigMap全局配置。

Ingress注解示例：

apiVersion:networking.k8s.io/v1
kind:Ingress
metadata:
name:cafe-ingress
annotations:
 nginx.ingress.kubernetes.io/ssl-redirect:"true"
 nginx.ingress.kubernetes.io/force-ssl-redirect:"true"
 nginx.ingress.kubernetes.io/limit-rps:"100"
 nginx.ingress.kubernetes.io/limit-connections:"50"
 nginx.ingress.kubernetes.io/proxy-body-size:"50m"
 nginx.ingress.kubernetes.io/proxy-connect-timeout:"30"
 nginx.ingress.kubernetes.io/proxy-read-timeout:"60"
 nginx.ingress.kubernetes.io/proxy-send-timeout:"60"
 nginx.ingress.kubernetes.io/websocket-services:"ws-service"
 nginx.ingress.kubernetes.io/use-regex:"true"
spec:
ingressClassName:nginx
rules:
 -host:cafe.example.com
  http:
   paths:
    -path:/tea
     pathType:Exact
     backend:
      service:
       name:tea-svc
       port:
        number:80
    -path:/coffee
     pathType:Prefix
     backend:
      service:
       name:coffee-svc
       port:
        number:80

常用注解說(shuō)明：

2.3 全局配置ConfigMap

通過(guò)ConfigMap可以設(shè)置Nginx的全局默認(rèn)配置：

apiVersion:v1
kind:ConfigMap
metadata:
name:nginx-configuration
namespace:ingress-nginx
data:
proxy-body-size:"50m"
proxy-connect-timeout:"30"
proxy-read-timeout:"60"
proxy-send-timeout:"60"
use-forwarded-headers:"true"
compute-full-forwarded-for:"true"
use-proxy-protocol:"false"
enable-underscores-in-headers:"true"
large-client-header-buffers:"4 16k"
client-header-buffer-size:"4k"
keep-alive:"75"
keep-alive-requests:"1000"
upstream-keepalive-connections:"50"
upstream-keepalive-timeout:"60"
upstream-keepalive-requests:"10000"
enable-brotli:"true"
enable-gzip:"true"
gzip-level:"6"
gzip-types:"application/json application/javascript application/xml text/css text/html text/javascript"

2.4 TLS配置

Nginx Ingress Controller支持多種TLS配置方式。

簡(jiǎn)單TLS配置：

apiVersion:networking.k8s.io/v1
kind:Ingress
metadata:
name:tls-ingress
spec:
ingressClassName:nginx
tls:
 -hosts:
   -example.com
   -www.example.com
  secretName:example-tls
rules:
 -host:example.com
  http:
   paths:
    -path:/
     pathType:Prefix
     backend:
      service:
       name:frontend
       port:
        number:80

自簽名證書(shū)創(chuàng)建：

# 創(chuàng)建私鑰
openssl genrsa -out tls.key 2048

# 創(chuàng)建證書(shū)
openssl req -new -x509 -key tls.key -out tls.crt -days 365 
  -subj"/CN=example.com/O=MyOrg"

# 創(chuàng)建Secret
kubectl create secret tls example-tls 
  --cert=tls.crt 
  --key=tls.key

# 或使用kubectl直接創(chuàng)建
kubectl create secret tls example-tls 
  --cert=tls.crt 
  --key=tls.key 
  --dry-run=client -o yaml | kubectl apply -f -

2.5 金絲雀發(fā)布

Nginx Ingress Controller支持基于權(quán)重和Header的金絲雀發(fā)布：

# 主版本Ingress
apiVersion:networking.k8s.io/v1
kind:Ingress
metadata:
name:main-ingress
spec:
ingressClassName:nginx
rules:
 -host:app.example.com
  http:
   paths:
    -path:/
     pathType:Prefix
     backend:
      service:
       name:main-service
       port:
        number:80
---
# 金絲雀Ingress
apiVersion:networking.k8s.io/v1
kind:Ingress
metadata:
name:canary-ingress
annotations:
 nginx.ingress.kubernetes.io/canary:"true"
 nginx.ingress.kubernetes.io/canary-weight:"30"
spec:
ingressClassName:nginx
rules:
 -host:app.example.com
  http:
   paths:
    -path:/
     pathType:Prefix
     backend:
      service:
       name:canary-service
       port:
        number:80

基于Header的金絲雀：

apiVersion:networking.k8s.io/v1
kind:Ingress
metadata:
name:header-canary-ingress
annotations:
 nginx.ingress.kubernetes.io/canary:"true"
 nginx.ingress.kubernetes.io/canary-by-header:"X-Canary"
 nginx.ingress.kubernetes.io/canary-by-header-value:"always"
spec:
ingressClassName:nginx
rules:
 -host:app.example.com
  http:
   paths:
    -path:/
     pathType:Prefix
     backend:
      service:
       name:canary-service
       port:
        number:80

2.6 限流配置

Nginx Ingress Controller提供多維度限流功能：

apiVersion:networking.k8s.io/v1
kind:Ingress
metadata:
name:rate-limit-ingress
annotations:
 nginx.ingress.kubernetes.io/limit-rps:"100"
 nginx.ingress.kubernetes.io/limit-rpm:"1000"
 nginx.ingress.kubernetes.io/limit-connections:"50"
 nginx.ingress.kubernetes.io/limit-burst-multiplier:"5"
spec:
ingressClassName:nginx
rules:
 -host:api.example.com
  http:
   paths:
    -path:/
     pathType:Prefix
     backend:
      service:
       name:api-service
       port:
        number:80

全局限流通過(guò)ConfigMap配置：

apiVersion:v1
kind:ConfigMap
metadata:
name:nginx-configuration
namespace:ingress-nginx
data:
limit_req_zone:"$binary_remote_addr zone=api:10m rate=10r/s"
limit_conn_zone:"$binary_remote_addr zone=addr:10m"
proxy-limit-rate:"0"

第三章 Traefik

3.1 架構(gòu)與原理

Traefik是一個(gè)現(xiàn)代云原生的反向代理和負(fù)載均衡器，采用動(dòng)態(tài)配置機(jī)制。與傳統(tǒng)Nginx不同，Traefik的配置變更不需要重載進(jìn)程，它通過(guò)監(jiān)聽(tīng)Kubernetes API或Consul等配置中心實(shí)時(shí)更新路由規(guī)則。

Traefik的架構(gòu)由多個(gè)組件構(gòu)成：Provider負(fù)責(zé)從不同來(lái)源獲取配置；Router負(fù)責(zé)將請(qǐng)求匹配到相應(yīng)的服務(wù)；Middleware在請(qǐng)求到達(dá)服務(wù)前進(jìn)行修改，如認(rèn)證、限流、重試等；Service將請(qǐng)求轉(zhuǎn)發(fā)到實(shí)際的后端。

Traefik Ingress Controller的部署：

apiVersion:apps/v1
kind:Deployment
metadata:
name:traefik
namespace:ingress
spec:
replicas:2
selector:
 matchLabels:
  app:traefik
template:
 metadata:
  labels:
   app:traefik
 spec:
  serviceAccountName:traefik-ingress-controller
  containers:
   -name:traefik
    image:traefik:v3.0.4
    args:
     ---api.insecure
     ---accesslog
     ---entrypoints.http.address=:80
     ---entrypoints.https.address=:443
     ---providers.kubernetesingress
     ---providers.kubernetescrd
     ---log.level=INFO
     ---metrics.prometheus
    ports:
     -name:http
      containerPort:80
     -name:https
      containerPort:443
     -name:admin
      containerPort:8080
    livenessProbe:
     httpGet:
      path:/ping
      port:8080
     initialDelaySeconds:10
     periodSeconds:10
    readinessProbe:
     httpGet:
      path:/ping
      port:8080
     periodSeconds:5
    resources:
     requests:
      cpu:100m
      memory:128Mi
     limits:
      cpu:500m
      memory:512Mi

3.2 IngressRoute資源

Traefik提供了自己的CRD——IngressRoute，它比Kubernetes原生Ingress更強(qiáng)大，支持更多配置選項(xiàng)：

apiVersion:traefik.io/v1alpha1
kind:IngressRoute
metadata:
name:demo-ingressroute
namespace:default
spec:
entryPoints:
 -web
 -websecure
routes:
 -match:Host(`demo.example.com`)&&PathPrefix(`/api`)
  kind:Rule
  services:
   -name:api-service
    port:80
  middlewares:
   -name:strip-api-prefix
   -name:rate-limit
 -match:Host(`demo.example.com`)&&PathPrefix(`/`)
  kind:Rule
  services:
   -name:frontend-service
    port:80
tls:
 secretName:demo-tls-cert

3.3 Middleware中間件

Traefik的Middleware是強(qiáng)大的請(qǐng)求處理組件，可以在路由層面添加認(rèn)證、重試、重定向、限流等功能。

基本認(rèn)證中間件：

apiVersion:traefik.io/v1alpha1
kind:Middleware
metadata:
name:basic-auth
namespace:default
spec:
basicAuth:
 secret:basic-auth-secret
---
apiVersion:v1
kind:Secret
metadata:
name:basic-auth-secret
namespace:default
type:Opaque
stringData:
users:|
  admin:$apr1$H6uskkkW$IgXLP6ewTrSuBkTrqE8wj/

IP白名單中間件：

apiVersion:traefik.io/v1alpha1
kind:Middleware
metadata:
name:ip-whitelist
namespace:default
spec:
ipWhiteList:
 sourceRange:
  -"10.0.0.0/8"
  -"192.168.1.0/24"
 ipStrategy:
  depth:0

限流中間件：

apiVersion:traefik.io/v1alpha1
kind:Middleware
metadata:
name:rate-limit
namespace:default
spec:
rateLimit:
 average:100
 burst:50
 period:1s

重試中間件：

apiVersion:traefik.io/v1alpha1
kind:Middleware
metadata:
name:retry
namespace:default
spec:
retry:
 attempts:3
 initialInterval:100ms

重定向中間件：

apiVersion:traefik.io/v1alpha1
kind:Middleware
metadata:
name:https-redirect
namespace:default
spec:
redirectScheme:
 scheme:https
 permanent:true

StripPrefix中間件：

apiVersion:traefik.io/v1alpha1
kind:Middleware
metadata:
name:strip-prefix
namespace:default
spec:
stripPrefix:
 prefixes:
  -/api/v1
  -/static

3.4 動(dòng)態(tài)服務(wù)發(fā)現(xiàn)

Traefik支持多種Provider，除了Kubernetes Ingress和CRD外，還支持Consul、Etcd、ZooKeeper等外部配置源。

Consul Provider配置示例：

# 在Consul中注冊(cè)服務(wù)
curl -X PUT http://consul:8500/v1/agent/service/register 
  -d'{
   "ID": "web-1",
   "Name": "web",
   "Address": "10.0.1.100",
   "Port": 8080,
   "Check": {
    "HTTP": "http://10.0.1.100:8080/health",
    "Interval": "10s"
   }
  }'

Traefik Provider配置：

apiVersion:v1
kind:ConfigMap
metadata:
name:traefik-config
namespace:ingress
data:
providers-consul-catalog:|
  address: consul:8500
  endpoint: http://consul:8500
  prefix: traefik
  requireConsistent: true
  refreshInterval: 10s
  exposedByDefault: false

3.5 TCP服務(wù)支持

Traefik不僅支持HTTP/HTTPS，還支持TCP和UDP服務(wù)：

apiVersion:traefik.io/v1alpha1
kind:IngressRouteTCP
metadata:
name:mysql-ingressroute
namespace:default
spec:
entryPoints:
 -mysql
routes:
 -match:HostSNI(`mysql.example.com`)
  services:
   -name:mysql-service
    port:3306
tls:
 passthrough:true
---
apiVersion:v1
kind:Service
metadata:
name:mysql-service
namespace:default
spec:
type:ClusterIP
ports:
 -name:mysql
  port:3306
  targetPort:3306

TCP EntryPoint配置：

apiVersion:v1
kind:ConfigMap
metadata:
name:traefik-config
namespace:ingress
data:
servers-transport:|
  [serversTransport]
   insecureSkipVerify = true
entrypoints:|
  [entryPoints]
   [entryPoints.mysql]
    address = ":3306"
   [entryPoints.redis]
    address = ":6379"

第四章 Envoy Proxy

4.1 架構(gòu)與原理

Envoy是Lyft開(kāi)發(fā)的開(kāi)源邊緣和服務(wù)代理，專(zhuān)為云原生架構(gòu)設(shè)計(jì)。Envoy的架構(gòu)基于xDS協(xié)議，通過(guò)監(jiān)聽(tīng)配置源動(dòng)態(tài)更新路由規(guī)則，無(wú)需重啟進(jìn)程。

Envoy的核心概念包括：Listener監(jiān)聽(tīng)入口流量；Route配置路由規(guī)則；Cluster定義后端服務(wù)；Endpoint是具體的Pod地址；Filter是請(qǐng)求處理鏈；Health Check用于檢測(cè)后端服務(wù)健康狀態(tài)。

Envoy在Kubernetes中的部署通常通過(guò)Contour或Istio等項(xiàng)目實(shí)現(xiàn)。Contour是最常用的Envoy Ingress Controller實(shí)現(xiàn)。

4.2 Contour部署

Contour是Envoy的Kubernetes Ingress Controller實(shí)現(xiàn)，由Heptio（現(xiàn)VMware）開(kāi)發(fā)。

apiVersion:apps/v1
kind:DaemonSet
metadata:
name:contour
namespace:projectcontour
spec:
replicas:2
selector:
 matchLabels:
  app:contour
template:
 metadata:
  labels:
   app:contour
 spec:
  containers:
   -name:contour
    image:ghcr.io/projectcontour/contour:v1.28.2
    command:
     -contour
     -serve
     ---xds-address=0.0.0.0
     ---xds-port=8001
     ---envoy-http-port=8080
     ---envoy-https-port=8443
     ---config-path=/config/contour.yaml
    ports:
     -name:xds
      containerPort:8001
      protocol:TCP
     -name:http
      containerPort:8080
      protocol:TCP
     -name:https
      containerPort:8443
      protocol:TCP
    livenessProbe:
     httpGet:
      path:/healthz
      port:8001
     initialDelaySeconds:5
     periodSeconds:10
    readinessProbe:
     httpGet:
      path:/healthz
      port:8001
     periodSeconds:5
  -name:envoy
   image:ghcr.io/projectcontour/contour:v1.28.2
   command:
    -envoy
    --c/config/envoy.json
    ---service-clusterprojectcontour
    ---service-node$(NODE_NAME)
   env:
    -name:NODE_NAME
     valueFrom:
      fieldRef:
       fieldPath:spec.nodeName
   ports:
    -name:http
     containerPort:8080
     protocol:TCP
    -name:https
     containerPort:8443
     protocol:TCP

Contour也支持Gateway API，這是Kubernetes網(wǎng)絡(luò)的新標(biāo)準(zhǔn)：

apiVersion:gateway.networking.k8s.io/v1
kind:Gateway
metadata:
name:prod-gateway
namespace:ingress
spec:
gatewayClassName:contour
listeners:
 -name:http
  port:80
  protocol:HTTP
  allowedRoutes:
   namespaces:
    from:Same
 -name:https
  port:443
  protocol:HTTPS
  allowedRoutes:
   namespaces:
    from:Same
  tls:
   mode:Terminate
   certificateRefs:
    -name:demo-cert
---
apiVersion:gateway.networking.k8s.io/v1
kind:HTTPRoute
metadata:
name:demo-route
namespace:default
spec:
parentRefs:
 -name:prod-gateway
  namespace:ingress
hostnames:
 -"demo.example.com"
rules:
 -backendRefs:
   -name:demo-service
    port:80

4.3 HTTPProxy資源

Contour擴(kuò)展了Kubernetes Ingress，提供了更強(qiáng)大的HTTPProxy CRD：

apiVersion:projectcontour.io/v1
kind:HTTPProxy
metadata:
name:demo-proxy
namespace:default
spec:
virtualhost:
 fqdn:demo.example.com
 cors:
  allow-credentials:true
  allow-headers:
   -X-Custom-Header
  allow-methods:
   -GET
   -POST
   -PUT
   -DELETE
  allow-origin:
   -"https://allowed.example.com"
  max-age:"86400"
routes:
 -conditions:
   -prefix:/api
  services:
   -name:api-service
    port:80
    healthcheck:
     path:/health
     interval:10s
     timeout:5s
     unhealthythreshold:3
     healthythreshold:2
  loadBalancerPolicy:
   strategy:WeightedLeastRequest
  retryPolicy:
   retryOn:gateway-error,connect-failure,reset
   numRetries:3
   perTryTimeout:10s
 -conditions:
   -prefix:/
  services:
   -name:frontend-service
    port:80
  rateLimit:
   global:
    descriptors:
     -entries:
       -key:remote_addr
        rateLimitValue:
         requests:100
         unit:minute
tcpproxy:
 services:
  -name:tcp-service
   port:9000
   weight:1

4.4 限流配置

Envoy支持多種限流策略：全局限流基于令牌桶算法，所有服務(wù)共享限流配額；本地限流在每個(gè)Envoy實(shí)例獨(dú)立執(zhí)行；基于請(qǐng)求屬性的限流可以組合多個(gè)維度。

全局限流示例：

apiVersion:projectcontour.io/v1
kind:TLSPolicy
metadata:
name:ratelimit-policy
namespace:default
spec:
limits:
 -units:second
  requests:100
  condition:
   -requestHeader:
     headerName:X-Forwarded-For
     count:1

本地限流注解：

apiVersion:v1
kind:Service
metadata:
name:api-service
annotations:
 projectcontour.io/lb-num-retries:"3"
 projectcontour.io/retry-on:"gateway-error,connect-failure,reset"
 projectcontour.io/per-try-timeout:"10s"
spec:
ports:
 -port:80
  targetPort:8080
selector:
 app:api

4.5 健康檢查與負(fù)載均衡

Envoy支持主動(dòng)和被動(dòng)兩種健康檢查方式。

主動(dòng)健康檢查配置：

apiVersion:projectcontour.io/v1
kind:HTTPProxy
metadata:
name:healthcheck-proxy
namespace:default
spec:
virtualhost:
 fqdn:api.example.com
routes:
 -services:
   -name:api-service
    port:80
    healthcheck:
     path:/healthz
     interval:10s
     timeout:5s
     expected-status:"200-299"
     unhealthythreshold:3
     healthythreshold:2

負(fù)載均衡策略：

apiVersion:projectcontour.io/v1
kind:HTTPProxy
metadata:
name:lb-proxy
namespace:default
spec:
virtualhost:
 fqdn:app.example.com
routes:
 -loadBalancerPolicy:
   strategy:Random
  services:
   -name:app-service
    port:80

支持的負(fù)載均衡策略：

第五章 三者綜合對(duì)比

5.1 性能對(duì)比

性能是Ingress Controller的重要考量因素，以下是基于公開(kāi)測(cè)試數(shù)據(jù)的對(duì)比。

原生Nginx性能最優(yōu)，它的請(qǐng)求處理延遲最低，吞吐量最高。這得益于Nginx多年的性能優(yōu)化和事件驅(qū)動(dòng)架構(gòu)。

Traefik v3相比v2版本性能提升顯著，但在高并發(fā)場(chǎng)景下仍略遜于Nginx。Traefik的動(dòng)態(tài)配置能力帶來(lái)了少量性能開(kāi)銷(xiāo)，但對(duì)于中低流量應(yīng)用影響很小。

Envoy采用C++實(shí)現(xiàn)，性能表現(xiàn)優(yōu)秀。Envoy的優(yōu)勢(shì)在于其可擴(kuò)展性，支持在請(qǐng)求處理鏈中添加多個(gè)Filter而不顯著影響性能。

實(shí)測(cè)數(shù)據(jù)參考（單實(shí)例，8核CPU，16GB內(nèi)存）：

5.2 功能特性對(duì)比

5.3 配置復(fù)雜度對(duì)比

Nginx Ingress的配置相對(duì)直觀，對(duì)于熟悉Nginx的運(yùn)維工程師來(lái)說(shuō)學(xué)習(xí)曲線(xiàn)平緩。大量在線(xiàn)資源和企業(yè)實(shí)踐經(jīng)驗(yàn)使得問(wèn)題解決相對(duì)容易。

Traefik的配置最簡(jiǎn)潔，聲明式的IngressRoute和Middleware設(shè)計(jì)使得配置易于理解。它的自動(dòng)服務(wù)發(fā)現(xiàn)減少了手動(dòng)配置工作。

Envoy的功能最強(qiáng)大，但配置也最復(fù)雜。Contour在一定程度上簡(jiǎn)化了Envoy的配置，但高級(jí)特性仍需要理解Envoy的架構(gòu)。對(duì)于團(tuán)隊(duì)技術(shù)實(shí)力較強(qiáng)、需要細(xì)粒度控制的場(chǎng)景，Envoy是理想選擇。

5.4 生態(tài)與社區(qū)對(duì)比

Nginx Ingress擁有最成熟的企業(yè)生態(tài)，大量生產(chǎn)環(huán)境案例和完善的文檔。Nginx公司提供商業(yè)支持Nginx Ingress Controller Plus。

Traefik由Traefik Labs維護(hù)，社區(qū)活躍，文檔質(zhì)量高。Traefik企業(yè)版提供更多企業(yè)級(jí)特性。

Envoy是Envoyproxy基金會(huì)和CNCF項(xiàng)目的重要組成部分，也是Istio服務(wù)網(wǎng)格的數(shù)據(jù)平面。Envoy在服務(wù)網(wǎng)格領(lǐng)域占據(jù)主導(dǎo)地位，選擇Envoy可以獲得整個(gè)云原生生態(tài)系統(tǒng)的好處。

第六章 選型建議與最佳實(shí)踐

6.1 場(chǎng)景化選型

中小規(guī)模集群（Pod數(shù)量小于1000）

推薦方案：Nginx Ingress Controller或Traefik

原因：配置簡(jiǎn)單，文檔完善，生態(tài)成熟。Nginx的性能足夠應(yīng)對(duì)中低流量，業(yè)務(wù)場(chǎng)景不需要復(fù)雜的金絲雀發(fā)布和流量管理。

大規(guī)模高流量集群

推薦方案：Nginx Ingress Controller或Envoy

原因：需要關(guān)注性能和穩(wěn)定性。Nginx在簡(jiǎn)單場(chǎng)景下性能最優(yōu)，Envoy在需要細(xì)粒度流量管理時(shí)更有優(yōu)勢(shì)。

需要精細(xì)流量管理的場(chǎng)景

推薦方案：Envoy（Contour）

原因：Envoy的xDS協(xié)議和豐富的Filter鏈支持復(fù)雜的流量管理需求，如熔斷、重試、限流、故障注入等。

服務(wù)網(wǎng)格集成

推薦方案：Envoy

原因：Istio、Linkerd等主流服務(wù)網(wǎng)格都使用Envoy作為數(shù)據(jù)平面。選擇Envoy Ingress Controller可以與現(xiàn)有服務(wù)網(wǎng)格技術(shù)棧保持一致。

6.2 部署架構(gòu)建議

生產(chǎn)環(huán)境建議使用DaemonSet模式部署Ingress Controller，確保每個(gè)節(jié)點(diǎn)都有入口：

apiVersion:apps/v1
kind:DaemonSet
metadata:
name:nginx-ingress
namespace:ingress-nginx
spec:
selector:
 matchLabels:
  app:nginx-ingress
template:
 metadata:
  labels:
   app:nginx-ingress
 spec:
  hostNetwork:true
  dnsPolicy:ClusterFirstWithHostNet
  containers:
   -name:controller
    image:registry.k8s.io/ingress-nginx/controller:v1.9.4
    args:
     -/nginx-ingress-controller
     ---configmap=$(POD_NAMESPACE)/nginx-configuration
     ---report-node-internal-ip-address
    env:
     -name:POD_NAME
      valueFrom:
       fieldRef:
        fieldPath:metadata.name
     -name:POD_NAMESPACE
      valueFrom:
       fieldRef:
        fieldPath:metadata.namespace
    ports:
     -name:http
      hostPort:80
      containerPort:80
     -name:https
      hostPort:443
      containerPort:443

高可用部署建議：

apiVersion:autoscaling/v2
kind:HorizontalPodAutoscaler
metadata:
name:nginx-ingress
namespace:ingress-nginx
spec:
scaleTargetRef:
 apiVersion:apps/v1
 kind:Deployment
 name:nginx-ingress-controller
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

6.3 安全配置

TLS配置最佳實(shí)踐：

apiVersion:networking.k8s.io/v1
kind:Ingress
metadata:
name:secure-ingress
annotations:
 nginx.ingress.kubernetes.io/ssl-redirect:"true"
 nginx.ingress.kubernetes.io/hsts-max-age:"31536000"
 nginx.ingress.kubernetes.io/hsts-include-subdomains:"true"
 nginx.ingress.kubernetes.io/proxy-buffer-size:"128k"
 nginx.ingress.kubernetes.io/server-snippet:|
   add_header X-Frame-Options "SAMEORIGIN" always;
   add_header X-Content-Type-Options "nosniff" always;
   add_header X-XSS-Protection "1; mode=block" always;
   add_header Referrer-Policy "no-referrer-when-downgrade" always;
spec:
ingressClassName:nginx
tls:
 -hosts:
   -example.com
  secretName:example-tls
rules:
 -host:example.com
  http:
   paths:
    -path:/
     pathType:Prefix
     backend:
      service:
       name:frontend
       port:
        number:80

6.4 監(jiān)控配置

Prometheus指標(biāo)暴露是生產(chǎn)環(huán)境的必備能力：

apiVersion:v1
kind:Service
metadata:
name:nginx-ingress-controller-metrics
namespace:ingress-nginx
annotations:
 prometheus.io/port:"10254"
 prometheus.io/scrape:"true"
spec:
ports:
 -name:metrics
  port:10254
  targetPort:metrics
selector:
 app:nginx-ingress
---
apiVersion:monitoring.coreos.com/v1
kind:ServiceMonitor
metadata:
name:nginx-ingress
namespace:monitoring
spec:
selector:
 matchLabels:
  app:nginx-ingress
endpoints:
 -port:metrics
  interval:15s
namespaceSelector:
 matchNames:
  -ingress-nginx

Grafana儀表板推薦使用公開(kāi)社區(qū)儀表板：

Nginx Ingress Controller: ID 9614

Traefik: ID 10162

Envoy: ID 13261

第七章 排障指南

7.1 查看Ingress Controller狀態(tài)

# 查看Controller日志
kubectl logs -n ingress-nginx -l app=nginx-ingress -f

# 查看最近事件
kubectl get events -n ingress-nginx --sort-by='.lastTimestamp'

# 查看Controller配置
kubectlexec-n ingress-nginx deploy/nginx-ingress-controller -- 
  nginx-ingress-controller --show-config

7.2 常見(jiàn)問(wèn)題與解決

問(wèn)題1：Ingress無(wú)法訪(fǎng)問(wèn)，返回404

排查步驟：

# 1. 檢查Ingress資源是否存在
kubectl get ingress -n default

# 2. 檢查Ingress地址配置
kubectl describe ingress demo-ingress -n default

# 3. 檢查Endpoints是否就緒
kubectl get endpoints -n default

# 4. 檢查Service配置
kubectl get svc -n default

# 5. 檢查Pod是否運(yùn)行正常
kubectl get pods -n default -l app=demo

# 6. 檢查DNS解析
kubectlexec-ittest-pod -- nslookup demo-service

問(wèn)題2：證書(shū)錯(cuò)誤或TLS不工作

# 檢查Secret是否存在
kubectl get secret demo-tls -n default

# 驗(yàn)證證書(shū)內(nèi)容
kubectl get secret demo-tls -n default -o jsonpath='{.data.tls.crt}'| base64 -d | openssl x509 -text -noout

# 檢查證書(shū)過(guò)期時(shí)間
kubectl get secret demo-tls -n default -o jsonpath='{.data.tls.crt}'| base64 -d | openssl x509 -enddate -noout

# 檢查Cert-Manager狀態(tài)（如使用）
kubectl get certificate -n default
kubectl describe certificate demo-cert -n default

問(wèn)題3：限流未生效

# 檢查限流注解是否正確配置
kubectl describe ingress rate-limit-ingress -n default

# 查看Controller日志中的限流信息
kubectl logs -n ingress-nginx -l app=nginx-ingress | grep -ilimit

# 驗(yàn)證ConfigMap中的限流配置
kubectl get configmap nginx-configuration -n ingress-nginx -o yaml

7.3 性能問(wèn)題排查

# 查看資源使用情況
kubectl top pods -n ingress-nginx

# 檢查連接數(shù)
kubectlexec-it nginx-ingress-controller-xxx -n ingress-nginx -- 
  wget -qO- http://localhost:10254/status

# 查看upstream狀態(tài)（Nginx Ingress）
kubectlexec-it nginx-ingress-controller-xxx -n ingress-nginx -- 
  cat /etc/nginx/nginx.conf | grep -A 50"upstream"

7.4 日志分析

Nginx Ingress訪(fǎng)問(wèn)日志格式配置：

apiVersion:v1
kind:ConfigMap
metadata:
name:nginx-configuration
namespace:ingress-nginx
data:
log-format-upstream:|
  '$remote_addr $remote_user [$time_local] "$request" '
  '$status $body_bytes_sent "$http_referer" '
  '"$http_user_agent" "$http_x_forwarded_for" '
  '$request_time $upstream_response_time $upstream_connect_time '
  'route=$upstream_route'

分析日志發(fā)現(xiàn)慢請(qǐng)求：

# 提取慢請(qǐng)求日志
kubectl logs -n ingress-nginx -l app=nginx-ingress | 
  awk'{if($NF > 1) print}'| sort -k9 -nr | head -20

結(jié)語(yǔ)

Nginx Ingress Controller、Traefik和Envoy是當(dāng)前最主流的三種Ingress Controller方案。它們各有優(yōu)勢(shì)：Nginx性能最優(yōu)、生態(tài)最成熟；Traefik配置最簡(jiǎn)潔、動(dòng)態(tài)能力最強(qiáng)；Envoy功能最豐富、與服務(wù)網(wǎng)格集成最緊密。

選型時(shí)應(yīng)綜合考慮團(tuán)隊(duì)技術(shù)棧、性能需求、功能需求和運(yùn)維能力。對(duì)于大多數(shù)場(chǎng)景，Nginx Ingress Controller是穩(wěn)妥的選擇；對(duì)于追求配置簡(jiǎn)潔和動(dòng)態(tài)能力的團(tuán)隊(duì)，Traefik值得考慮；對(duì)于需要精細(xì)流量控制和計(jì)劃使用服務(wù)網(wǎng)格的場(chǎng)景，Envoy是理想選擇。

無(wú)論選擇哪種方案，都應(yīng)該建立完善的監(jiān)控告警體系，制定 Ingress Controller的運(yùn)維流程，并定期進(jìn)行故障演練，確保在生產(chǎn)環(huán)境中能夠快速響應(yīng)和恢復(fù)。

參考資料

Kubernetes Ingress官方文檔：https://kubernetes.io/docs/concepts/services-networking/ingress/

Nginx Ingress Controller官方文檔：https://kubernetes.github.io/ingress-nginx/

Traefik官方文檔：https://doc.traefik.io/traefik/

Contour官方文檔：https://projectcontour.io/docs/

Envoy官方文檔：https://www.envoyproxy.io/docs/envoy/latest/