不懂envoyfilter也敢说精通istio系列-ratelimit-istio ratelimit完全手册

迈不过友情╰ 2022-08-28 09:59 111阅读 0赞

#   欢迎关注我的公众号： #

![watermark_type_ZHJvaWRzYW5zZmFsbGJhY2s_shadow_50_text_Q1NETiBAaHhwamF2YTE_size_8_color_FFFFFF_t_70_g_se_x_16][]

目前刚开始写一个月，一共写了18篇原创文章，文章目录如下：

[istio多集群探秘，部署了50次多集群后我得出的结论][istio_50]

[istio多集群链路追踪，附实操视频][istio]

[istio防故障利器，你知道几个,istio新手不要读，太难！][istio_istio]

[istio业务权限控制，原来可以这么玩][istio 1]

[istio实现非侵入压缩，微服务之间如何实现压缩][istio 2]

[不懂envoyfilter也敢说精通istio系列-http-rbac-不要只会用AuthorizationPolicy配置权限][envoyfilter_istio_-http-rbac-_AuthorizationPolicy]

[不懂envoyfilter也敢说精通istio系列-02-http-corsFilter-不要只会vs][envoyfilter_istio_-02-http-corsFilter-_vs]

[不懂envoyfilter也敢说精通istio系列-03-http-csrf filter-再也不用再代码里写csrf逻辑了][envoyfilter_istio_-03-http-csrf filter-_csrf]

[不懂envoyfilter也敢说精通istio系列http-jwt\_authn-不要只会RequestAuthorization][envoyfilter_istio_http-jwt_authn-_RequestAuthorization]

[不懂envoyfilter也敢说精通istio系列-05-fault-filter-故障注入不止是vs][envoyfilter_istio_-05-fault-filter-_vs]

[不懂envoyfilter也敢说精通istio系列-06-http-match-配置路由不只是vs][envoyfilter_istio_-06-http-match-_vs]

[不懂envoyfilter也敢说精通istio系列-07-负载均衡配置不止是dr][envoyfilter_istio_-07-_dr]

[不懂envoyfilter也敢说精通istio系列-08-连接池和断路器][envoyfilter_istio_-08-]

[不懂envoyfilter也敢说精通istio系列-09-http-route filter][envoyfilter_istio_-09-http-route filter]

[不懂envoyfilter也敢说精通istio系列-network filter-redis proxy][envoyfilter_istio_-network filter-redis proxy]

[不懂envoyfilter也敢说精通istio系列-network filter-HttpConnectionManager][envoyfilter_istio_-network filter-HttpConnectionManager]

[不懂envoyfilter也敢说精通istio系列-ratelimit-istio ratelimit完全手册][envoyfilter_istio_-ratelimit-istio ratelimit]

# 1什么是ratelimit #

ratelimit是限速的意思，主要作用是限值流量，防止因系统过载而崩溃。istio限速有两大类，一个是本地限速，另一个是全局限速。本地限速是在envoy内部提供一种令牌桶限速得功能，全局限速需要访问外部限速服务。按生效位置分可分为virtualhost级别的限速和route级别的限速。按协议分可分为http的限速和tcp的限速。按服务位置分可分为mesh内部限速和对外部请求的限速。按集群分可分为单集群限速和多集群限速。

# 2配置 #

extensions.filters.http.ratelimit.v3.RateLimit：

{
      "domain": "...",对应ratelimit配置文件中的域名
      "stage": "...",stage配置和action中的stage必须匹配，默认0
      "request_type": "...", internal, external or both默认both
      "timeout": "{...}",ratelimit服务请求的超时时间，默认20ms
      "failure_mode_deny": "...",如果ratelimit服务请求失败，则返回错误
      "rate_limited_as_resource_exhausted": "...",RESOURCE_EXHAUSTED code代替UNAVAILABLE code
      "rate_limit_service": "{...}",指定外部限速服务
      "enable_x_ratelimit_headers": "...",定义ratelimit头版本
      "disable_x_envoy_ratelimited_header": "..."禁用x-envoy-ratelimited头
    }

rate\_limit\_service：

{
      "grpc_service": "{...}",配置ratelimit服务
      "transport_api_version": "..."版本
    }

grpc\_service：

{
      "envoy_grpc": "{...}",envoy grpc客户端
      "google_grpc": "{...}",google grpc客户端
      "timeout": "{...}",grpc请求超时时间
      "initial_metadata": []额外的元数据信息
    }

envoy\_grpc：

{
      "cluster_name": "...",集群名称
      "authority": "..." :authority头的值，默认cluster_name
    }

google\_grpc：

{
      "target_uri": "...",目标uri
      "channel_credentials": "{...}",秘钥
      "call_credentials": [],调用秘钥
      "stat_prefix": "...",stat前缀
      "credentials_factory_name": "...",秘钥工厂名称
      "config": "{...}",额外配置
      "per_stream_buffer_limit_bytes": "{...}",每个流的缓存限值，默认1MiB
      "channel_args": "{...}"通道参数
    }

transport\_api\_version：

xDS API and non-xDS services version. This is used to describe both resource and transport protocol versions (in distinct configuration fields).

*  AUTO
    
    *(DEFAULT)* ⁣When not specified, we assume v2, to ease migration to Envoy’s stable API versioning. If a client does not support v2 (e.g. due to deprecation), this is an invalid value.
 *  V2
    
    ⁣Use xDS v2 API.
 *  V3
    
    ⁣Use xDS v3 API.

enable\_x\_ratelimit\_headers：

Defines the version of the standard to use for X-RateLimit headers.

*  OFF
    
    *(DEFAULT)* ⁣X-RateLimit headers disabled.
 *  DRAFT\_VERSION\_03
    
    ⁣Use [draft RFC Version 03][].

extensions.filters.network.local\_ratelimit.v3.LocalRateLimit

{
      "stat_prefix": "...",stat前缀
      "token_bucket": "{...}",令牌桶规则
      "runtime_enabled": "{...}"启用百分比
    }

token\_bucket：

{
      "max_tokens": "...",最大令牌数
      "tokens_per_fill": "{...}",每次填充的令牌数
      "fill_interval": "{...}"填充间隔
    }

runtime\_enabled：

{
      "default_value": "{...}",百分比
      "runtime_key": "..."运行时key
    }

config.route.v3.RateLimit

{
      "stage": "{...}",阶段号，必须和ratelimit过滤器匹配
      "disable_key": "...",禁用的key
      "actions": [],动作
      "limit": "{...}"动态元数据
    }

actions：

{
      "source_cluster": "{...}",源集群动作
      "destination_cluster": "{...}",目标集群动作
      "request_headers": "{...}",请求头动作
      "remote_address": "{...}",远程地址动作
      "generic_key": "{...}",通用key动作
      "header_value_match": "{...}",头匹配动作
      "dynamic_metadata": "{...}",动态元数据动作
      "metadata": "{...}",元数据动作
      "extension": "{...}"扩展动作
    }

limit：

{
      "dynamic_metadata": "{...}"动态元数据
    }

extensions.filters.network.ratelimit.v3.RateLimit

{
      "stat_prefix": "...",
      "domain": "...",
      "descriptors": [],
      "timeout": "{...}",
      "failure_mode_deny": "...",
      "rate_limit_service": "{...}"
    }

# 3实战 #

## 3.1.1http ##

## 3.1.1.1单集群 ##

### 3.1.1.1.1集群内服务限流 ###

3.1.1.1.1.1本地限流

cat <<EOF > envoyfilter-local-rate-limit.yaml
    apiVersion: networking.istio.io/v1alpha3
    kind: EnvoyFilter
    metadata:
      name: filter-local-ratelimit-svc
    spec:
      workloadSelector:
        labels:
          app: productpage
      configPatches:
        - applyTo: HTTP_FILTER
          match:
            listener:
              filterChain:
                filter:
                  name: "envoy.filters.network.http_connection_manager"
          patch:
            operation: INSERT_BEFORE
            value:
              name: envoy.filters.http.local_ratelimit
              typed_config:
                "@type": type.googleapis.com/udpa.type.v1.TypedStruct
                type_url: type.googleapis.com/envoy.extensions.filters.http.local_ratelimit.v3.LocalRateLimit
                value:
                  stat_prefix: http_local_rate_limiter
                  token_bucket:
                    max_tokens: 10
                    tokens_per_fill: 10
                    fill_interval: 60s
                  filter_enabled:
                    runtime_key: local_rate_limit_enabled
                    default_value:
                      numerator: 100
                      denominator: HUNDRED
                  filter_enforced:
                    runtime_key: local_rate_limit_enforced
                    default_value:
                      numerator: 100
                      denominator: HUNDRED
                  response_headers_to_add:
                    - append: false
                      header:
                        key: x-local-rate-limit
                        value: 'true'
    EOF
    
    kubectl apply -f envoyfilter-local-rate-limit.yaml -n istio

说明：本地限流需要通过EnvoyFilter来实现，他不会请求外部服务，在envoy内部实现支持，是一个令牌桶的算法。http filter的名称必须是envoy.filters.http.local\_ratelimit，type和typeurl是固定的，stat\_prefix可以随便改，表示生成stat的指标前缀。token\_bucket配置令牌桶，max\_tokens表示最大令牌数量，tokens\_per\_fill表示每次填充的令牌数量，fill\_interval表示填充令牌的间隔。filter\_enabled表示启用但不是强制，filter\_enforced表示强制，可以配置百分比。response\_headers\_to\_add修改响应头信息，append为false表示修改，true表示添加。runtime\_key 运行时的key，具体有啥用不清楚。

执行压测：

[root@node01 45]# go-stress-testing -c 10 -n 10000 -u http://192.168.229.134:30945/productpage
    
     开始启动  并发数:10 请求数:10000 请求参数: 
    request:
     form:http 
     url:http://192.168.229.134:30945/productpage 
     method:GET 
     headers:map[] 
     data: 
     verify:statusCode 
     timeout:30s 
     debug:false 
    
    
    
    ─────┬───────┬───────┬───────┬────────┬────────┬────────┬────────┬────────┬────────┬────────
     耗时│ 并发数│ 成功数│ 失败数│   qps  │最长耗时│最短耗时│平均耗时│下载字节│字节每秒│ 错误码
    ─────┼───────┼───────┼───────┼────────┼────────┼────────┼────────┼────────┼────────┼────────
       1s│      7│      2│    761│    2.94│  124.68│    1.98│ 3406.97│  21,476│  21,470│200:2;429:761
       2s│     10│      5│   1636│    2.55│ 1788.46│    1.98│ 3928.11│  52,771│  26,383│200:5;429:1636
       3s│     10│      5│   2962│    1.70│ 1788.46│    1.04│ 5871.68│  76,639│  25,545│200:5;429:2962
       4s│     10│      5│   4459│    1.28│ 1788.46│    1.04│ 7810.78│ 103,585│  25,896│200:5;429:4459

429 Too Many Requests (太多请求)

当你需要限制客户端请求某个服务的数量，也就是限制请求速度时，该状态码就会非常有用

清理：

kubectl delete envoyfilter filter-local-ratelimit-svc -n istio

3.1.1.1.1.2全局限流

部署ratelimit

1创建cm

cat << EOF > ratelimit-config.yaml
    apiVersion: v1
    kind: ConfigMap
    metadata:
      name: ratelimit-config
    data:
      config.yaml: |
        domain: productpage-ratelimit
        descriptors:
          - key: PATH
            value: "/productpage"
            rate_limit:
              unit: minute
              requests_per_unit: 1
          - key: PATH
            rate_limit:
              unit: minute
              requests_per_unit: 100
    EOF
    
    kubectl apply -f ratelimit-config.yaml -n istio

说明: 这个configmap是限速服务用到的配置文件，他是envoy v3版本的限速格式。domain是域名，他会在envoyfilter中被引用，descriptors的PATH,表示请求的路径可以有多个值，rate\_limit配置限速配额，这里productpage配了1分钟1个请求，其他url是1分钟100个请求

2创建限速服务deployment

cat << EOF > ratelimit-deploy.yaml
    apiVersion: v1
    kind: Service
    metadata:
      name: redis
      labels:
        app: redis
    spec:
      ports:
      - name: redis
        port: 6379
      selector:
        app: redis
    ---
    apiVersion: apps/v1
    kind: Deployment
    metadata:
      name: redis
    spec:
      replicas: 1
      selector:
        matchLabels:
          app: redis
      template:
        metadata:
          labels:
            app: redis
        spec:
          containers:
          - image: redis:alpine
            imagePullPolicy: Always
            name: redis
            ports:
            - name: redis
              containerPort: 6379
          restartPolicy: Always
          serviceAccountName: ""
    ---
    apiVersion: v1
    kind: Service
    metadata:
      name: ratelimit
      labels:
        app: ratelimit
    spec:
      ports:
      - name: http-port
        port: 8080
        targetPort: 8080
        protocol: TCP
      - name: grpc-port
        port: 8081
        targetPort: 8081
        protocol: TCP
      - name: http-debug
        port: 6070
        targetPort: 6070
        protocol: TCP
      selector:
        app: ratelimit
    ---
    apiVersion: apps/v1
    kind: Deployment
    metadata:
      name: ratelimit
    spec:
      replicas: 1
      selector:
        matchLabels:
          app: ratelimit
      strategy:
        type: Recreate
      template:
        metadata:
          labels:
            app: ratelimit
        spec:
          containers:
          - image: envoyproxy/ratelimit:6f5de117 # 2021/01/08
            imagePullPolicy: Always
            name: ratelimit
            command: ["/bin/ratelimit"]
            env:
            - name: LOG_LEVEL
              value: debug
            - name: REDIS_SOCKET_TYPE
              value: tcp
            - name: REDIS_URL
              value: redis:6379
            - name: USE_STATSD
              value: "false"
            - name: RUNTIME_ROOT
              value: /data
            - name: RUNTIME_SUBDIRECTORY
              value: ratelimit
            ports:
            - containerPort: 8080
            - containerPort: 8081
            - containerPort: 6070
            volumeMounts:
            - name: config-volume
              mountPath: /data/ratelimit/config/config.yaml
              subPath: config.yaml
          volumes:
          - name: config-volume
            configMap:
              name: ratelimit-config
    EOF
    
    kubectl apply -f ratelimit-deploy.yaml -n istio

创建了redis，和官方的一个ratelimit服务。

3创建envoy-filter

cat << EOF > envoyfilter-filter.yaml
    apiVersion: networking.istio.io/v1alpha3
    kind: EnvoyFilter
    metadata:
      name: filter-ratelimit
      namespace: istio-system
    spec:
      workloadSelector:
        # select by label in the same namespace
        labels:
          istio: ingressgateway
      configPatches:
        # The Envoy config you want to modify
        - applyTo: HTTP_FILTER
          match:
            context: GATEWAY
            listener:
              filterChain:
                filter:
                  name: "envoy.filters.network.http_connection_manager"
                  subFilter:
                    name: "envoy.filters.http.router"
          patch:
            operation: INSERT_BEFORE
            # Adds the Envoy Rate Limit Filter in HTTP filter chain.
            value:
              name: envoy.filters.http.ratelimit
              typed_config:
                "@type": type.googleapis.com/envoy.extensions.filters.http.ratelimit.v3.RateLimit
                # domain can be anything! Match it to the ratelimter service config
                domain: productpage-ratelimit
                failure_mode_deny: true
                rate_limit_service:
                  grpc_service:
                    envoy_grpc:
                      cluster_name: rate_limit_cluster
                    timeout: 10s
                  transport_api_version: V3
        - applyTo: CLUSTER
          match:
            cluster:
              service: ratelimit.istio.svc.cluster.local
          patch:
            operation: ADD
            # Adds the rate limit service cluster for rate limit service defined in step 1.
            value:
              name: rate_limit_cluster
              type: STRICT_DNS
              connect_timeout: 10s
              lb_policy: ROUND_ROBIN
              http2_protocol_options: {}
              load_assignment:
                cluster_name: rate_limit_cluster
                endpoints:
                - lb_endpoints:
                  - endpoint:
                      address:
                         socket_address:
                          address: ratelimit.istio.svc.cluster.local
                          port_value: 8081
    EOF
    
    kubectl apply -f envoyfilter-filter.yaml -n istio-system

这个envoyfilter作用在网关上面，配置了一个http过滤器envoy.filters.http.ratelimit，和一个cluster。http 过滤器的cluster地址指向cluster配置的地址，这里就是我们的ratelimit service所在的地址。domain是上面configmap的值一样，failure\_mode\_deny表示超过请求限值就拒绝，rate\_limit\_service配置ratelimit服务的地址（cluster），这里可以配置grpc类型的也可以配置http类型的。

4创建action envoyfilter

cat << EOF > envoyfilter-action.yaml 
    apiVersion: networking.istio.io/v1alpha3
    kind: EnvoyFilter
    metadata:
      name: filter-ratelimit-svc
      namespace: istio-system
    spec:
      workloadSelector:
        labels:
          istio: ingressgateway
      configPatches:
        - applyTo: VIRTUAL_HOST
          match:
            context: GATEWAY
            routeConfiguration:
              vhost:
                name: "*:80"
                route:
                  action: ANY
          patch:
            operation: MERGE
            # Applies the rate limit rules.
            value:
              rate_limits:
                - actions: # any actions in here
                  - request_headers:
                      header_name: ":path"
                      descriptor_key: "PATH"
    EOF
    
    kubectl apply -f envoyfilter-action.yaml  -n istio-system

这个envoyfilter作用在入口网关处，给80端口的虚拟主机配置了一个rate\_limits 动作，descriptor\_key用于选择在configmap里配置的key。

压测：

[root@node01 ~]# go-stress-testing -c 10 -n 10000 -u http://192.168.229.134:30945/productpage
    
     开始启动  并发数:10 请求数:10000 请求参数: 
    request:
     form:http 
     url:http://192.168.229.134:30945/productpage 
     method:GET 
     headers:map[] 
     data: 
     verify:statusCode 
     timeout:30s 
     debug:false 
    
    
    
    ─────┬───────┬───────┬───────┬────────┬────────┬────────┬────────┬────────┬────────┬────────
     耗时│ 并发数│ 成功数│ 失败数│   qps  │最长耗时│最短耗时│平均耗时│下载字节│字节每秒│ 错误码
    ─────┼───────┼───────┼───────┼────────┼────────┼────────┼────────┼────────┼────────┼────────
       1s│     10│      1│   1051│    1.01│   55.51│    3.70│ 9914.38│   4,183│   4,176│200:1;429:1051
       2s│     10│      1│   1629│    0.50│   55.51│    3.70│19807.86│   4,183│   2,090│200:1;429:1629
       3s│     10│      1│   2154│    0.34│   55.51│    3.70│29829.63│   4,183│   1,393│200:1;429:2154
       4s│     10│      1│   2662│    0.25│   55.51│    3.70│39823.69│   4,183│   1,045│200:1;429:2662
       5s│     10│      1│   3166│    0.20│   58.63│    3.70│49865.16│   4,183│     836│200:1;429:3166

清理：

kubectl delete cm ratelimit-config -n istio
    kubectl delete -f ratelimit-deploy.yaml -n istio
    kubectl delete envoyfilter filter-ratelimit -n istio-system
    kubectl delete envoyfilter filter-ratelimit-svc -n istio-system

### 3.1.1.1.2集群外服务限流 ###

3.1.1.1.2.1本地限流

cat << EOF > se-baidu.yaml
    apiVersion: networking.istio.io/v1alpha3
    kind: ServiceEntry
    metadata:
      name: baidu
    spec:
      hosts:
        - www.baidu.com
      ports:
        - number: 80
          name: http-port
          protocol: HTTP
      resolution: DNS
    EOF
    
    kubectl apply -f se-baidu.yaml -n istio

说明，创建了访问百度的service entry

cat <<EOF > envoyfilter-local-rate-limit-http-outside.yaml
    apiVersion: networking.istio.io/v1alpha3
    kind: EnvoyFilter
    metadata:
      name: filter-local-ratelimit-svc
    spec:
      workloadSelector:
        labels:
          app: ratings
      configPatches:
        - applyTo: HTTP_FILTER
          match:
            context: SIDECAR_OUTBOUND
            listener:
              filterChain:
                filter:
                  name: "envoy.filters.network.http_connection_manager"
                  subFilter:
                    name: "envoy.filters.http.router"
          patch:
            operation: INSERT_BEFORE
            value:
              name: envoy.filters.http.local_ratelimit
              typed_config:
                "@type": type.googleapis.com/udpa.type.v1.TypedStruct
                type_url: type.googleapis.com/envoy.extensions.filters.http.local_ratelimit.v3.LocalRateLimit
                value:
                  stat_prefix: http_local_rate_limiter
                  token_bucket:
                    max_tokens: 1
                    tokens_per_fill: 1
                    fill_interval: 60s
                  filter_enabled:
                    runtime_key: local_rate_limit_enabled
                    default_value:
                      numerator: 100
                      denominator: HUNDRED
                  filter_enforced:
                    runtime_key: local_rate_limit_enforced
                    default_value:
                      numerator: 100
                      denominator: HUNDRED
                  response_headers_to_add:
                    - append: false
                      header:
                        key: x-local-rate-limit
                        value: 'true'
    EOF
    
    kubectl apply -f envoyfilter-local-rate-limit-http-outside.yaml -n istio

说明：SIDECAR\_OUTBOUND表示对外发出请求起作用。本地限流需要通过EnvoyFilter来实现，他不会请求外部服务，在envoy内部实现支持，是一个令牌桶的算法。http filter的名称必须是envoy.filters.http.local\_ratelimit，type和typeurl是固定的，stat\_prefix可以随便改，表示生成stat的指标前缀。token\_bucket配置令牌桶，max\_tokens表示最大令牌数量，tokens\_per\_fill表示每次填充的令牌数量，fill\_interval表示填充令牌的间隔。filter\_enabled表示启用但不是强制，filter\_enforced表示强制，可以配置百分比。response\_headers\_to\_add修改响应头信息，append为false表示修改，true表示添加。runtime\_key 运行时的key，具体有啥用不清楚。

kubectl exec -it -n istio ratings-v2-mysql-vm-66dc56449d-lk6gv /bin/bash
    
    local_rate_limitednode@ratings-v2-mysql-vm-66dc56449d-lk6gv:/opt/microservices$ curl www.baidu.com -I
    HTTP/1.1 429 Too Many Requests
    x-local-rate-limit: true
    content-length: 18
    content-type: text/plain
    date: Fri, 17 Sep 2021 23:20:13 GMT
    server: envoy

进入ratings容器，对百度发请求，409错误，说明限流生效

清理：

kubectl delete se baidu -n istio
    kubectl delete envoyfilter filter-local-ratelimit-svc -n istio

3.1.1.1.2.2全局限流

部署ratelimit

1创建cm

cat << EOF > ratelimit-config-outside-http.yaml
    apiVersion: v1
    kind: ConfigMap
    metadata:
      name: ratelimit-config
    data:
      config.yaml: |
        domain: productpage-ratelimit
        descriptors:
          - key: PATH
            value: "/"
            rate_limit:
              unit: minute
              requests_per_unit: 1
          - key: PATH
            value: "/aa"
            rate_limit:
              unit: minute
              requests_per_unit: 1
          - key: PATH
            rate_limit:
              unit: minute
              requests_per_unit: 100
    EOF
    
    kubectl apply -f ratelimit-config-outside-http.yaml -n istio

说明: 这个configmap是限速服务用到的配置文件，他是envoy v3版本的限速格式。domain是域名，他会在envoyfilter中被引用，descriptors的PATH,表示请求的路径可以有多个值，rate\_limit配置限速配额，这里productpage配了1分钟1个请求，/aa一分钟1个请求，其他url是1分钟100个请求

2创建限速服务deployment

创建了redis，和官方的一个ratelimit服务。

创建访问百度的serviceentry

创建envoy-filter

cat << EOF > envoyfilter-filter-outside-http.yaml
    apiVersion: networking.istio.io/v1alpha3
    kind: EnvoyFilter
    metadata:
      name: filter-ratelimit
      namespace: istio
    spec:
      workloadSelector:
        # select by label in the same namespace
        labels:
          app: ratings
      configPatches:
        - applyTo: HTTP_FILTER
          match:
            context: SIDECAR_OUTBOUND
            listener:
              filterChain:
                filter:
                  name: "envoy.filters.network.http_connection_manager"
                  subFilter:
                    name: "envoy.filters.http.router"
          patch:
            operation: INSERT_BEFORE
            # Adds the Envoy Rate Limit Filter in HTTP filter chain.
            value:
              name: envoy.filters.http.ratelimit
              typed_config:
                "@type": type.googleapis.com/envoy.extensions.filters.http.ratelimit.v3.RateLimit
                # domain can be anything! Match it to the ratelimter service config
                domain: productpage-ratelimit
                failure_mode_deny: true
                rate_limit_service:
                  grpc_service:
                    envoy_grpc:
                      cluster_name: rate_limit_cluster
                    timeout: 10s
                  transport_api_version: V3
        - applyTo: CLUSTER
          match:
            cluster:
              service: ratelimit.istio.svc.cluster.local
          patch:
            operation: ADD
            # Adds the rate limit service cluster for rate limit service defined in step 1.
            value:
              name: rate_limit_cluster
              type: STRICT_DNS
              connect_timeout: 10s
              lb_policy: ROUND_ROBIN
              http2_protocol_options: {}
              load_assignment:
                cluster_name: rate_limit_cluster
                endpoints:
                - lb_endpoints:
                  - endpoint:
                      address:
                         socket_address:
                          address: ratelimit.istio.svc.cluster.local
                          port_value: 8081
    EOF
    
    kubectl apply -f envoyfilter-filter-outside-http.yaml -n istio

这个envoyfilter作用在ratings上面，SIDECAR\_OUTBOUND作用的对外流量上面，配置了一个http过滤器envoy.filters.http.ratelimit，和一个cluster。http 过滤器的cluster地址指向cluster配置的地址，这里就是我们的ratelimit service所在的地址。domain是上面configmap的值一样，failure\_mode\_deny表示超过请求限值就拒绝，rate\_limit\_service配置ratelimit服务的地址（cluster），这里可以配置grpc类型的也可以配置http类型的。

4创建action envoyfilter

cat << EOF > envoyfilter-action-outside-http.yaml 
    apiVersion: networking.istio.io/v1alpha3
    kind: EnvoyFilter
    metadata:
      name: filter-ratelimit-svc
      namespace: istio
    spec:
      workloadSelector:
        labels:
          app: ratings
      configPatches:
        - applyTo: VIRTUAL_HOST
          match:
            context: SIDECAR_OUTBOUND
            routeConfiguration:
              vhost:
                name: "www.baidu.com:80"
                route:
                  action: ANY
          patch:
            operation: MERGE
            # Applies the rate limit rules.
            value:
              rate_limits:
                - actions: # any actions in here
                  - request_headers:
                      header_name: ":path"
                      descriptor_key: "PATH"
    EOF
    
    kubectl apply -f envoyfilter-action-outside-http.yaml   -n istio

host是我们配置的百度的地址（www.baidu.com:80），这个envoyfilter作用在ratings处，给80端口的虚拟主机配置了一个rate\_limits 动作，descriptor\_key用于选择在configmap里配置的key。

kubectl exec -it -n istio ratings-v2-mysql-vm-66dc56449d-lk6gv /bin/bash
    
    node@ratings-v2-mysql-vm-66dc56449d-lk6gv:/opt/microservices$ curl www.baidu.com/ -I
    HTTP/1.1 429 Too Many Requests
    x-envoy-ratelimited: true
    date: Fri, 17 Sep 2021 23:51:33 GMT
    server: envoy
    transfer-encoding: chunked

进入rating容器，向百度发请求，409错误，说明限流成功

清理：

kubectl delete cm ratelimit-config -n istio
    kubectl delete -f ratelimit-deploy.yaml -n istio
    kubectl delete envoyfilter filter-ratelimit -n istio
    kubectl delete envoyfilter filter-ratelimit-svc -n istio

## 3.1.1.2多集群 ##

### 3.1.1.2.1准备集群 ###

![watermark_type_ZHJvaWRzYW5zZmFsbGJhY2s_shadow_50_text_Q1NETiBAaHhwamF2YTE_size_20_color_FFFFFF_t_70_g_se_x_16][]

这里多集群安装我们不在展开，不懂得可以看我之前的文章。

集群1
    128,129,130
    集群2
    131,132,133
    
    两个网络联通
    128。129.130
    route add -net 172.21.1.0 netmask 255.255.255.0 gw 192.168.229.131
    route add -net 172.21.2.0 netmask 255.255.255.0 gw 192.168.229.133
    route add -net 172.21.0.0 netmask 255.255.255.0 gw 192.168.229.132
    route add -net 10.69.0.0 netmask 255.255.0.0 gw 192.168.229.131
    
    131,132，133
    route add -net 172.20.0.0 netmask 255.255.255.0 gw 192.168.229.128
    route add -net 172.20.1.0 netmask 255.255.255.0 gw 192.168.229.129
    route add -net 172.20.2.0 netmask 255.255.255.0 gw 192.168.229.130
    route add -net 10.68.0.0 netmask 255.255.0.0 gw 192.168.229.128
    
    
    cluster1:
    生成istio安装operator文件
    cat <<EOF > cluster1.yaml
    apiVersion: install.istio.io/v1alpha1
    kind: IstioOperator
    spec:
      profile: demo
      values:
        global:
          meshID: mesh1
          multiCluster:
            clusterName: cluster1
          network: network1
      meshConfig:
        accessLogFile: /dev/stdout
        enableTracing: true
      components:
        egressGateways:
        - name: istio-egressgateway
          enabled: true
    EOF
    
    
    生成istio安装operator文件
    cat <<EOF > cluster2.yaml
    apiVersion: install.istio.io/v1alpha1
    kind: IstioOperator
    spec:
      profile: demo
      values:
        global:
          meshID: mesh1
          multiCluster:
            clusterName: cluster2
          network: network1
      meshConfig:
        accessLogFile: /dev/stdout
        enableTracing: true
      components:
        egressGateways:
        - name: istio-egressgateway
          enabled: true
    EOF
    
    把部署文件传到cluster2
    scp cluster2.yaml root@192.168.229.131:/root
    
    
    cluster1:
    生成访问apiserver secret
     istioctl x create-remote-secret --name=cluster1  --server=https://192.168.229.128:6443 > remote-secret-cluster1.yaml
     传输secret到cluster2
    scp remote-secret-cluster1.yaml root@192.168.229.131:/root
    
    cluster2
    生成访问apiserver secret
     istioctl x create-remote-secret --name=cluster2  --server=https://192.168.229.131:6443 > remote-secret-cluster2.yaml
      传输secret到cluster2
     scp remote-secret-cluster2.yaml root@192.168.229.128:/root
     
     cluster1
     应用secret
     kubectl apply -f remote-secret-cluster2.yaml
     
     部署集群
     istioctl install  -f cluster1.yaml
      
      
      cluster2:
      应用secret
      kubectl apply -f remote-secret-cluster1.yaml
      部署集群
    istioctl install  -f cluster2.yaml
    
     cluster1: 
      重启pod
     kubectl rollout restart deploy -n istio
     kubectl rollout restart deploy -n istio-system
     
     cluster2:
       重启pod
     kubectl rollout restart deploy -n istio
     kubectl rollout restart deploy -n istio-system

清理：

cluster1:
    
    kubectl delete secret istio-remote-secret-cluster2 -n istio-system
    istioctl x uninstall -f cluster1.yaml
    
    reboot
    
    cluster2:
    
    kubectl delete secret istio-remote-secret-cluster1 -n istio-system
    istioctl x uninstall -f cluster2.yaml
    
    reboot

### 3.1.1.2.2集群内本地限流 ###

多集群只演示集群内本地限流其他是一样的

cat <<EOF > envoyfilter-local-rate-limit-multi-http.yaml
    apiVersion: networking.istio.io/v1alpha3
    kind: EnvoyFilter
    metadata:
      name: filter-local-ratelimit-svc
    spec:
      workloadSelector:
        labels:
          app: productpage
      configPatches:
        - applyTo: HTTP_FILTER
          match:
            listener:
              filterChain:
                filter:
                  name: "envoy.filters.network.http_connection_manager"
          patch:
            operation: INSERT_BEFORE
            value:
              name: envoy.filters.http.local_ratelimit
              typed_config:
                "@type": type.googleapis.com/udpa.type.v1.TypedStruct
                type_url: type.googleapis.com/envoy.extensions.filters.http.local_ratelimit.v3.LocalRateLimit
                value:
                  stat_prefix: http_local_rate_limiter
                  token_bucket:
                    max_tokens: 10
                    tokens_per_fill: 10
                    fill_interval: 60s
                  filter_enabled:
                    runtime_key: local_rate_limit_enabled
                    default_value:
                      numerator: 100
                      denominator: HUNDRED
                  filter_enforced:
                    runtime_key: local_rate_limit_enforced
                    default_value:
                      numerator: 100
                      denominator: HUNDRED
                  response_headers_to_add:
                    - append: false
                      header:
                        key: x-local-rate-limit
                        value: 'true'
    EOF
    
    kubectl apply -f envoyfilter-local-rate-limit-multi-http.yaml -n istio

开启压测：

[root@node01 45]# go-stress-testing -c 10 -n 10000 -u http://192.168.229.128:30363/productpage
    
     开始启动  并发数:10 请求数:10000 请求参数: 
    request:
     form:http 
     url:http://192.168.229.128:30363/productpage 
     method:GET 
     headers:map[] 
     data: 
     verify:statusCode 
     timeout:30s 
     debug:false 
    
    
    
    ─────┬───────┬───────┬───────┬────────┬────────┬────────┬────────┬────────┬────────┬────────
     耗时│ 并发数│ 成功数│ 失败数│   qps  │最长耗时│最短耗时│平均耗时│下载字节│字节每秒│ 错误码
    ─────┼───────┼───────┼───────┼────────┼────────┼────────┼────────┼────────┼────────┼────────
       1s│      0│      0│      0│    0.00│    0.00│    0.00│    0.00│        │        │
       2s│      7│     16│      6│   15.25│ 1453.38│    4.56│  655.73│  71,950│  35,974│200:16;429:6
       3s│      7│     17│      6│   14.30│ 1453.38│    4.56│  699.44│  76,133│  25,376│200:17;429:6
       4s│     10│     34│     24│   14.30│ 3207.96│    2.71│  699.46│ 154,262│  38,559│200:34;429:24
       5s│     10│     78│     68│   16.03│ 3207.96│    2.71│  623.67│ 370,054│  74,009│200:78;429:68
       6s│     10│    160│    150│   26.98│ 3207.96│    2.71│  370.66│ 770,420│ 128,402│200:160;429:150
       7s│     10│    238│    228│   34.53│ 3207.96│    2.71│  289.60│1,148,994│ 164,131│200:238;429:228

发现多集群和单集群的配置是一样，同样生效

## 3.2.2tcp ##

### 3.2.2.1单集群 ###

3.2.2.1.1集群内服务限流

3.2.2.1.1.1本地限流

部署mysql

cat << EOF > mysql.yaml
    apiVersion: v1
    kind: Secret
    metadata:
      name: mysql-credentials
    type: Opaque
    data:
      rootpasswd: cGFzc3dvcmQ=
    ---
    apiVersion: v1
    kind: Service
    metadata:
      name: mysqldb
      labels:
        app: mysqldb
        service: mysqldb
    spec:
      ports:
      - port: 3306
        name: tcp
      selector:
        app: mysqldb
    ---
    apiVersion: apps/v1
    kind: Deployment
    metadata:
      name: mysqldb-v1
      labels:
        app: mysqldb
        version: v1
    spec:
      replicas: 1
      selector:
        matchLabels:
          app: mysqldb
          version: v1
      template:
        metadata:
          labels:
            app: mysqldb
            version: v1
        spec:
          containers:
          - name: mysqldb
            image: docker.io/istio/examples-bookinfo-mysqldb:1.16.2
            imagePullPolicy: IfNotPresent
            ports:
            - containerPort: 3306
            env:
              - name: MYSQL_ROOT_PASSWORD
                valueFrom:
                  secretKeyRef:
                    name: mysql-credentials
                    key: rootpasswd
            args: ["--default-authentication-plugin","mysql_native_password"]
            volumeMounts:
            - name: var-lib-mysql
              mountPath: /var/lib/mysql
          volumes:
          - name: var-lib-mysql
            emptyDir: {}
    EOF
    
    kubectl apply -f mysql.yaml -n istio

说明：部署了mysql服务，ratings获取数据时会请求这个服务

部署mysql版ratings

cat << EOF > ratings-mysql.yaml
    apiVersion: apps/v1
    kind: Deployment
    metadata:
      name: ratings-v2-mysql
      labels:
        app: ratings
        version: v2-mysql
    spec:
      replicas: 1
      selector:
        matchLabels:
          app: ratings
          version: v2-mysql
      template:
        metadata:
          labels:
            app: ratings
            version: v2-mysql
        spec:
          containers:
          - name: ratings
            image: docker.io/istio/examples-bookinfo-ratings-v2:1.16.2
            imagePullPolicy: IfNotPresent
            env:
              # ratings-v2 will use mongodb as the default db backend.
              # if you would like to use mysqldb then you can use this file
              # which sets DB_TYPE = 'mysql' and the rest of the parameters shown
              # here and also create the # mysqldb service using bookinfo-mysql.yaml
              # NOTE: This file is mutually exclusive to bookinfo-ratings-v2.yaml
              - name: DB_TYPE
                value: "mysql"
              - name: MYSQL_DB_HOST
                value: mysqldb
              - name: MYSQL_DB_PORT
                value: "3306"
              - name: MYSQL_DB_USER
                value: root
              - name: MYSQL_DB_PASSWORD
                value: password
            ports:
            - containerPort: 9080
            securityContext:
              runAsUser: 1000
    EOF
    
    kubectl apply -f ratings-mysql.yaml -n istio

部署了mysql版的ratings，是指了env变量。

cat <<EOF > envoyfilter-local-rate-limit-mysql-inside.yaml
    apiVersion: networking.istio.io/v1alpha3
    kind: EnvoyFilter
    metadata:
      name: filter-local-ratelimit-svc
    spec:
      workloadSelector:
        labels:
          app: mysqldb
          version: v1
      configPatches:
        - applyTo: NETWORK_FILTER
          match:
            listener:
              portNumber: 3306
              filterChain:
                filter:
                  name: "envoy.filters.network.tcp_proxy"
          patch:
            operation: INSERT_BEFORE
            value:
              name: envoy.filters.network.local_ratelimit
              typed_config:
                "@type": type.googleapis.com/udpa.type.v1.TypedStruct
                type_url: type.googleapis.com/envoy.extensions.filters.network.local_ratelimit.v3.LocalRateLimit
                value:
                  stat_prefix: tcp_local_rate_limiter
                  token_bucket:
                    max_tokens: 1
                    tokens_per_fill: 1
                    fill_interval: 60s
                  runtime_enabled:
                    runtime_key: tcp_rate_limit_enabled
                    default_value: true
    EOF
    
    kubectl apply -f envoyfilter-local-rate-limit-mysql-inside.yaml -n istio

注意这里applyTo的是NETWORK\_FILTER，因为mysql是tcp服务，不是http服务，filter的名字是envoy.filters.network.local\_ratelimit，type\_url也是固定的不要写错。token\_bucket配置了访问限速的令牌数量及其填充速度。我们设置的filter在envoy.filters.network.tcp\_proxy前面，所以是INSERT\_BEFORE。

![watermark_type_ZHJvaWRzYW5zZmFsbGJhY2s_shadow_50_text_Q1NETiBAaHhwamF2YTE_size_20_color_FFFFFF_t_70_g_se_x_16 1][]

清理：

kubectl delete -f  mysql.yaml -n istio
    kubectl delete -f ratings-mysql.yaml -n istio
    kubectl delete envoyfilter filter-local-ratelimit-svc -n istio

3.2.2.1.1.2全局限流

部署mysql

wechat/envoyfilter/ratelimit/tcp/inner/global/mysql.yaml

部署mysql版ratings

部署ratelimit

1创建cm

cat << EOF > ratelimit-config-mysql-inside.yaml
    apiVersion: v1
    kind: ConfigMap
    metadata:
      name: ratelimit-config
    data:
      config.yaml: |
        domain: mysql-ratelimit
        descriptors:
          - key: test
            value: "test"
            rate_limit:
              unit: minute
              requests_per_unit: 1
          - key: test
            rate_limit:
              unit: minute
              requests_per_unit: 10
    EOF
    
    kubectl apply -f ratelimit-config-mysql-inside.yaml -n istio

2创建限速服务deployment

创建了redis，和官方的一个ratelimit服务。

3创建envoy-filter

cat << EOF > envoyfilter-filter-mysql-inside.yaml
    apiVersion: networking.istio.io/v1alpha3
    kind: EnvoyFilter
    metadata:
      name: filter-ratelimit
      namespace: istio
    spec:
      workloadSelector:
        labels:
          app: mysqldb
          version: v1
      configPatches:
        - applyTo: NETWORK_FILTER
          match:
            listener:
              portNumber: 3306
              filterChain:
                filter:
                  name: "envoy.filters.network.tcp_proxy"
          patch:
            operation: INSERT_BEFORE
            value:
              name: envoy.filters.network.ratelimit
              typed_config:
                "@type": type.googleapis.com/envoy.extensions.filters.network.ratelimit.v3.RateLimit
                domain: mysql-ratelimit
                failure_mode_deny: true
                stat_prefix: mysql_ratelimit
                descriptors:
                - entries:
                  - key: test
                    value: test
                  limit: 
                    requests_per_unit: 2
                    unit: MINUTE
                rate_limit_service:
                  grpc_service:
                    envoy_grpc:
                      cluster_name: rate_limit_cluster
                    timeout: 10s
                  transport_api_version: V3
        - applyTo: CLUSTER
          match:
            cluster:
              service: ratelimit.istio.svc.cluster.local
          patch:
            operation: ADD
            value:
              name: rate_limit_cluster
              type: STRICT_DNS
              connect_timeout: 10s
              lb_policy: ROUND_ROBIN
              http2_protocol_options: {}
              load_assignment:
                cluster_name: rate_limit_cluster
                endpoints:
                - lb_endpoints:
                  - endpoint:
                      address:
                         socket_address:
                          address: ratelimit.istio.svc.cluster.local
                          port_value: 8081
    EOF
    
    kubectl apply -f envoyfilter-filter-mysql-inside.yaml -n istio

node@ratings-v2-mysql-565f8fd887-8hp9s:/opt/microservices$ curl ratings.istio.svc:9080/ratings/0
    {"id":0,"ratings":{"Reviewer1":5,"Reviewer2":4}}node@ratings-v2-mysql-565f8fd887-8hp9s:/opt/microservices$ curl ratings.istio.svc:9080/ratings/0
    {"error":"could not connect to ratings database"}node@ratings-v2-mysql-565f8fd887-8hp9s:/opt/microservices$ curl ratings.istio.svc:9080/ratings/0
    {"error":"could not connect to ratings database"}node@ratings-v2-mysql-565f8fd887-8hp9s:/opt/microservices$

清理：

kubectl delete -f mysql.yaml -n istio
    kubectl delete -f ratings-mysql.yaml -n istio
    kubectl delete -f envoyfilter-filter-mysql-inside.yaml -n istio
    kubectl delete -f ratelimit-deploy.yaml -n istio
    kubectl delete -f ratelimit-config-mysql-inside.yaml -n istio

3.2.2.1.2集群外服务限流

3.2.2.1.2.1本地限流

部署rating-v2

cat << EOF > bookinfo-ratings-v2-mysql-vm.yaml
    apiVersion: apps/v1
    kind: Deployment
    metadata:
      name: ratings-v2-mysql-vm
      labels:
        app: ratings
        version: v2-mysql-vm
    spec:
      replicas: 1
      selector:
        matchLabels:
          app: ratings
          version: v2-mysql-vm
      template:
        metadata:
          labels:
            app: ratings
            version: v2-mysql-vm
        spec:
          containers:
          - name: ratings
            image: docker.io/istio/examples-bookinfo-ratings-v2:1.16.2
            imagePullPolicy: IfNotPresent
            env:
              # This assumes you registered your mysql vm as
              # istioctl register -n vm mysqldb 1.2.3.4 3306
              - name: DB_TYPE
                value: "mysql"
              - name: MYSQL_DB_HOST
                value: mysql.vm.demo
              - name: MYSQL_DB_PORT
                value: "3306"
              - name: MYSQL_DB_USER
                value: root
              - name: MYSQL_DB_PASSWORD
                value: root
            ports:
            - containerPort: 9080
            securityContext:
              runAsUser: 1000
    EOF
    
    
    kubectl apply -f bookinfo-ratings-v2-mysql-vm.yaml -n istio

在vm上部署mysql，这个略过，有需要文档的同学，可以加我微信，因为有点复杂

创建serviceentry

cat << EOF > se-mysql.yaml
    apiVersion: networking.istio.io/v1beta1
    kind: ServiceEntry
    metadata:
      name: mysql-se
    spec:
      hosts:
      - mysql.vm.demo
      addresses:
      - 192.168.229.12
      location: MESH_INTERNAL
      ports:
      - number: 3306
        name: mysql
        protocol: TCP
        targetPort: 3306
      resolution: STATIC
      workloadSelector:
        labels:
          app: mysql
          type: vm
    EOF
    
    kubectl apply -f se-mysql.yaml -n vm

这里创建了一个访问我们部署的虚拟机服务mysql的serviceentry

coredns配置加上解析记录

apiVersion: v1
    data:
      Corefile: |
        .:53 {
            errors
            health {
                lameduck 5s
            }
            hosts {
                192.168.229.11  httpd.vm.demo
                192.168.229.12  mysql.vm.demo
                fallthrough
            }
            ready
            kubernetes cluster.local in-addr.arpa ip6.arpa {
                pods insecure
                fallthrough in-addr.arpa ip6.arpa
                ttl 30
            }
            prometheus :9153
            forward . /etc/resolv.conf {
                max_concurrent 1000
            }
            cache 30
            reload
            loadbalance
        }
    kind: ConfigMap

添加192.168.229.12 mysql.vm.demo这一段

重启coredns

kubectl rollout restart -n kube-system deployment coredns

执行压测

go-stress-testing -c 1 -n 10000 -u  http://192.168.229.134:32688/productpage

[root@node01 ~]# go-stress-testing -c 1 -n 10000 -u  http://192.168.229.134:32688/productpage
    
     开始启动  并发数:1 请求数:10000 请求参数: 
    request:
     form:http 
     url:http://192.168.229.134:32688/productpage 
     method:GET 
     headers:map[] 
     data: 
     verify:statusCode 
     timeout:30s 
     debug:false 
    
    
    
    ─────┬───────┬───────┬───────┬────────┬────────┬────────┬────────┬────────┬────────┬────────
     耗时│ 并发数│ 成功数│ 失败数│   qps  │最长耗时│最短耗时│平均耗时│下载字节│字节每秒│ 错误码
    ─────┼───────┼───────┼───────┼────────┼────────┼────────┼────────┼────────┼────────┼────────
       1s│      1│     18│      0│   18.20│   91.36│   17.66│   54.94│  87,270│  87,264│200:18
       2s│      1│     37│      0│   19.11│   91.40│   13.13│   52.34│ 178,723│  89,351│200:37
       3s│      1│     54│      0│   18.42│   97.80│   13.13│   54.30│ 260,814│  86,928│200:54
       4s│      1│     72│      0│   18.22│   97.80│   13.13│   54.88│ 349,080│  87,258│200:72
       5s│      1│     90│      0│   18.17│  103.04│   12.98│   55.04│ 436,350│  87,264│200:90
       6s│      1│    111│      0│   18.59│  103.04│   12.98│   53.80│ 538,165│  89,686│200:111
       7s│      1│    132│      0│   18.91│  103.04│   12.98│   52.89│ 638,984│  91,279│200:132
       8s│      1│    150│      0│   18.88│  103.04│   12.30│   52.95│ 727,250│  90,905│200:150

没有限流之前压测都是200的返回结果

cat <<EOF > envoyfilter-local-rate-limit-mysql-vm-outside.yaml
    apiVersion: networking.istio.io/v1alpha3
    kind: EnvoyFilter
    metadata:
      name: filter-local-ratelimit-svc
    spec:
      workloadSelector:
        labels:
          app: ratings
      configPatches:
        - applyTo: NETWORK_FILTER
          match:
            context: SIDECAR_OUTBOUND
            listener:
              portNumber: 3306
              filterChain:
                filter:
                  name: "envoy.filters.network.tcp_proxy"
          patch:
            operation: INSERT_BEFORE
            value:
              name: envoy.filters.network.local_ratelimit
              typed_config:
                "@type": type.googleapis.com/udpa.type.v1.TypedStruct
                type_url: type.googleapis.com/envoy.extensions.filters.network.local_ratelimit.v3.LocalRateLimit
                value:
                  stat_prefix: tcp_local_rate_limiter
                  token_bucket:
                    max_tokens: 10
                    tokens_per_fill: 10
                    fill_interval: 60s
                  runtime_enabled:
                    runtime_key: tcp_rate_limit_enabled
                    default_value: true
    EOF
    
    kubectl apply -f envoyfilter-local-rate-limit-mysql-vm-outside.yaml -n istio

部署envoyfilter使限流生效，作用在ratings服务上面，而且是出口流量SIDECAR\_OUTBOUND，network filter名称必须事envoy.filters.network.local\_ratelimit，type \_url也是固定的。token\_bucket设置令牌桶。

![watermark_type_ZHJvaWRzYW5zZmFsbGJhY2s_shadow_50_text_Q1NETiBAaHhwamF2YTE_size_20_color_FFFFFF_t_70_g_se_x_16 2][]

这里无法连接数据库，说明数据库被限流了，ratings无法连接vm mysql服务。

清理：

kubectl delete envoyfilter filter-local-ratelimit-svc -n istio
    kubectl delete se mysql-se -n vm

3.2.2.1.2.2全局限流

部署rating-v2

在vm上部署mysql，这个略过，有需要文档的同学，可以加我微信，因为有点复杂

创建serviceentry

这里创建了一个访问我们部署的虚拟机服务mysql的serviceentry

coredns配置加上解析记录

添加192.168.229.12 mysql.vm.demo这一段

重启coredns

kubectl rollout restart -n kube-system deployment coredns

执行压测

go-stress-testing -c 1 -n 10000 -u  http://192.168.229.134:32688/productpage

没有限流之前压测都是200的返回结果

cat << EOF > envoyfilter-filter-mysql-outside.yaml
    apiVersion: networking.istio.io/v1alpha3
    kind: EnvoyFilter
    metadata:
      name: filter-ratelimit
      namespace: istio
    spec:
      workloadSelector:
        labels:
          app: ratings
      configPatches:
        - applyTo: NETWORK_FILTER
          match:
            context: SIDECAR_OUTBOUND
            listener:
              portNumber: 3306
              filterChain:
                filter:
                  name: "envoy.filters.network.tcp_proxy"
          patch:
            operation: INSERT_BEFORE
            value:
              name: envoy.filters.network.ratelimit
              typed_config:
                "@type": type.googleapis.com/envoy.extensions.filters.network.ratelimit.v3.RateLimit
                domain: mysql-ratelimit
                failure_mode_deny: true
                stat_prefix: mysql_ratelimit
                descriptors:
                - entries:
                  - key: test
                    value: test
                  limit: 
                    requests_per_unit: 2
                    unit: MINUTE
                rate_limit_service:
                  grpc_service:
                    envoy_grpc:
                      cluster_name: rate_limit_cluster
                    timeout: 10s
                  transport_api_version: V3
        - applyTo: CLUSTER
          match:
            cluster:
              service: ratelimit.istio.svc.cluster.local
          patch:
            operation: ADD
            value:
              name: rate_limit_cluster
              type: STRICT_DNS
              connect_timeout: 10s
              lb_policy: ROUND_ROBIN
              http2_protocol_options: {}
              load_assignment:
                cluster_name: rate_limit_cluster
                endpoints:
                - lb_endpoints:
                  - endpoint:
                      address:
                         socket_address:
                          address: ratelimit.istio.svc.cluster.local
                          port_value: 8081
    EOF
    
    kubectl apply -f envoyfilter-filter-mysql-outside.yaml -n istio

清理

kubectl delete envoyfilter filter-local-ratelimit-svc -n istio
    kubectl delete se mysql-se -n vm

3.2.2.2多集群

3.2.2.2.1集群准备

同上

3.2.2.2.2集群内本地限流

cluster1,cluster2部署mysql

说明：部署了mysql服务，ratings获取数据时会请求这个服务

cluster1,cluster2部署mysql版ratings

部署了mysql版的ratings，是指了env变量。

cat <<EOF > envoyfilter-local-rate-limit-mysql-inside.yaml
    apiVersion: networking.istio.io/v1alpha3
    kind: EnvoyFilter
    metadata:
      name: filter-local-ratelimit-svc
    spec:
      workloadSelector:
        labels:
          app: mysqldb
          version: v1
      configPatches:
        - applyTo: NETWORK_FILTER
          match:
            listener:
              portNumber: 3306
              filterChain:
                filter:
                  name: "envoy.filters.network.tcp_proxy"
          patch:
            operation: INSERT_BEFORE
            value:
              name: envoy.filters.network.local_ratelimit
              typed_config:
                "@type": type.googleapis.com/udpa.type.v1.TypedStruct
                type_url: type.googleapis.com/envoy.extensions.filters.network.local_ratelimit.v3.LocalRateLimit
                value:
                  stat_prefix: tcp_local_rate_limiter
                  token_bucket:
                    max_tokens: 1
                    tokens_per_fill: 1
                    fill_interval: 60s
                  runtime_enabled:
                    runtime_key: tcp_rate_limit_enabled
                    default_value: true
    EOF
    
    cluster1，cluster2：
    kubectl apply -f envoyfilter-local-rate-limit-mysql-inside.yaml -n istio

![watermark_type_ZHJvaWRzYW5zZmFsbGJhY2s_shadow_50_text_Q1NETiBAaHhwamF2YTE_size_20_color_FFFFFF_t_70_g_se_x_16 3][]

多集群集群内本地限流，需要在每个istiod里面增加ratelimit配置

清理：
    kubectl delete envoyfilter filter-local-ratelimit-svc -n istio

## 3.3不同限流动作 ##

### 3.3.1source\_cluster ###

部署ratelimit

1创建cm

cat << EOF > ratelimit-config.yaml
    apiVersion: v1
    kind: ConfigMap
    metadata:
      name: ratelimit-config
    data:
      config.yaml: |
        domain: productpage-ratelimit
        descriptors:
          - key: source_cluster
            value: "outbound|80||istio-ingressgateway.istio-system.svc.cluster.local"
            rate_limit:
              unit: minute
              requests_per_unit: 1
          - key: source_cluster
            rate_limit:
              unit: minute
              requests_per_unit: 10
    EOF
    
    kubectl apply -f ratelimit-config.yaml -n istio

2创建限速服务deployment

3创建envoy-filter

cat << EOF > envoyfilter-filter.yaml
    apiVersion: networking.istio.io/v1alpha3
    kind: EnvoyFilter
    metadata:
      name: filter-ratelimit
      namespace: istio
    spec:
      workloadSelector:
        # select by label in the same namespace
        labels:
          app: productpage
      configPatches:
        # The Envoy config you want to modify
        - applyTo: HTTP_FILTER
          match:
            context: SIDECAR_INBOUND
            listener:
              filterChain:
                filter:
                  name: "envoy.filters.network.http_connection_manager"
                  subFilter:
                    name: "envoy.filters.http.router"
          patch:
            operation: INSERT_BEFORE
            # Adds the Envoy Rate Limit Filter in HTTP filter chain.
            value:
              name: envoy.filters.http.ratelimit
              typed_config:
                "@type": type.googleapis.com/envoy.extensions.filters.http.ratelimit.v3.RateLimit
                # domain can be anything! Match it to the ratelimter service config
                domain: productpage-ratelimit
                failure_mode_deny: true
                rate_limit_service:
                  grpc_service:
                    envoy_grpc:
                      cluster_name: rate_limit_cluster
                    timeout: 10s
                  transport_api_version: V3
        - applyTo: CLUSTER
          match:
            cluster:
              service: ratelimit.istio.svc.cluster.local
          patch:
            operation: ADD
            # Adds the rate limit service cluster for rate limit service defined in step 1.
            value:
              name: rate_limit_cluster
              type: STRICT_DNS
              connect_timeout: 10s
              lb_policy: ROUND_ROBIN
              http2_protocol_options: {}
              load_assignment:
                cluster_name: rate_limit_cluster
                endpoints:
                - lb_endpoints:
                  - endpoint:
                      address:
                         socket_address:
                          address: ratelimit.istio.svc.cluster.local
                          port_value: 8081
    EOF
    
    kubectl apply -f envoyfilter-filter.yaml -n istio

4创建action envoyfilter

### 3.3.2destination\_cluster ###

部署ratelimit

1创建cm

cat << EOF > ratelimit-config.yaml
    apiVersion: v1
    kind: ConfigMap
    metadata:
      name: ratelimit-config
    data:
      config.yaml: |
        domain: productpage-ratelimit
        descriptors:
          - key: destination_cluster
            value: "inbound|9080||"
            rate_limit:
              unit: minute
              requests_per_unit: 1
          - key: destination_cluster
            rate_limit:
              unit: minute
              requests_per_unit: 10
    EOF
    
    kubectl apply -f ratelimit-config.yaml -n istio

2创建限速服务deployment

3创建envoy-filter

4创建action envoyfilter

[root@node01 ~]# go-stress-testing -n 1000000 -c 10 -u http://192.168.229.135:32688/productpage
    
     开始启动  并发数:10 请求数:1000000 请求参数: 
    request:
     form:http 
     url:http://192.168.229.135:32688/productpage 
     method:GET 
     headers:map[] 
     data: 
     verify:statusCode 
     timeout:30s 
     debug:false 
    
    
    
    ─────┬───────┬───────┬───────┬────────┬────────┬────────┬────────┬────────┬────────┬────────
     耗时│ 并发数│ 成功数│ 失败数│   qps  │最长耗时│最短耗时│平均耗时│下载字节│字节每秒│ 错误码
    ─────┼───────┼───────┼───────┼────────┼────────┼────────┼────────┼────────┼────────┼────────
       1s│     10│      1│    799│    1.01│   38.49│    5.84│ 9904.46│   4,183│   4,182│200:1;429:799
       2s│     10│      1│   1703│    0.50│   38.49│    5.31│19918.49│   4,183│   2,090│200:1;429:1703
       3s│     10│      1│   2553│    0.33│   38.49│    4.88│29894.78│   4,183│   1,394│200:1;429:2553
       4s│     10│      1│   3452│    0.25│   38.49│    4.88│39907.66│   4,183│   1,045│200:1;429:3452
       5s│     10│      1│   4339│    0.20│   38.49│    4.26│49863.27│   4,183│     836│200:1;429:4339
       6s│     10│      1│   5130│    0.17│   38.49│    4.26│59819.79│   4,183│     697│200:1;429:5130
       7s│     10│      1│   5945│    0.14│   38.49│    4.26│69888.47│   4,183│     597│200:1;429:5945
       8s│     10│      1│   6773│    0.13│   38.49│    4.26│79848.79│   4,183│     522│200:1;429:6772;500:1

### 3.3.3request\_headers ###

略

### 3.3.4remote\_address ###

部署ratelimit

1创建cm

cat << EOF > ratelimit-config.yaml
    apiVersion: v1
    kind: ConfigMap
    metadata:
      name: ratelimit-config
    data:
      config.yaml: |
        domain: productpage-ratelimit
        descriptors:
          - key: remote_address
            value: "172.20.1.0"
            rate_limit:
              unit: minute
              requests_per_unit: 1
          - key: remote_address
            rate_limit:
              unit: minute
              requests_per_unit: 10
    EOF
    
    kubectl apply -f ratelimit-config.yaml -n istio

2创建限速服务deployment

3创建envoy-filter

4创建action envoyfilter

### 3.3.5generic\_key ###

{
      "descriptor_value": "...",
      "descriptor_key": "..."
    }

部署ratelimit

1创建cm

cat << EOF > ratelimit-config.yaml
    apiVersion: v1
    kind: ConfigMap
    metadata:
      name: ratelimit-config
    data:
      config.yaml: |
        domain: productpage-ratelimit
        descriptors:
          - key: test
            value: "test"
            rate_limit:
              unit: minute
              requests_per_unit: 2
          - key: test
            rate_limit:
              unit: minute
              requests_per_unit: 10
    EOF
    
    kubectl apply -f ratelimit-config.yaml -n istio

2创建限速服务deployment

3创建envoy-filter

4创建action envoyfilter

cat << EOF > envoyfilter-action.yaml 
    apiVersion: networking.istio.io/v1alpha3
    kind: EnvoyFilter
    metadata:
      name: filter-ratelimit-svc
      namespace: istio
    spec:
      workloadSelector:
        labels:
          app: productpage
      configPatches:
        - applyTo: VIRTUAL_HOST
          match:
            context: SIDECAR_INBOUND
            routeConfiguration:
              vhost:
                name: "inbound|http|9080"
                route:
                  action: ANY
          patch:
            operation: MERGE
            # Applies the rate limit rules.
            value:
              rate_limits:
                - actions: 
                  - generic_key:
                      descriptor_key: test
                      descriptor_value: test
    EOF
    
    kubectl apply -f envoyfilter-action.yaml  -n istio

[root@node01 ~]# go-stress-testing -n 1000000 -c 10 -u http://192.168.229.135:32688/productpage
    
     开始启动  并发数:10 请求数:1000000 请求参数: 
    request:
     form:http 
     url:http://192.168.229.135:32688/productpage 
     method:GET 
     headers:map[] 
     data: 
     verify:statusCode 
     timeout:30s 
     debug:false 
    
    
    
    ─────┬───────┬───────┬───────┬────────┬────────┬────────┬────────┬────────┬────────┬────────
     耗时│ 并发数│ 成功数│ 失败数│   qps  │最长耗时│最短耗时│平均耗时│下载字节│字节每秒│ 错误码
    ─────┼───────┼───────┼───────┼────────┼────────┼────────┼────────┼────────┼────────┼────────
       1s│     10│      0│    789│    0.00│   27.99│    5.37│    0.00│        │        │429:789
       2s│     10│      0│   1624│    0.00│   27.99│    4.61│    0.00│        │        │429:1624
       3s│     10│      0│   2492│    0.00│   27.99│    4.61│    0.00│        │        │429:2492
       4s│     10│      0│   3335│    0.00│   27.99│    4.21│    0.00│        │        │429:3335
       5s│     10│      0│   4181│    0.00│   27.99│    4.21│    0.00│        │        │429:4181
       6s│     10│      0│   5026│    0.00│   27.99│    4.21│    0.00│        │        │429:5026
       7s│     10│      0│   5842│    0.00│   27.99│    4.21│    0.00│        │        │429:5842
       8s│     10│      0│   6721│    0.00│   27.99│    4.21│    0.00│        │        │429:6721
       9s│     10│      2│   7586│    0.22│   95.85│    4.21│44910.75│  10,362│   1,151│200:2;429:7586
      10s│     10│      2│   8444│    0.20│   95.85│    4.21│49900.13│  10,362│   1,036│200:2;429:8441;500:3
      11s│     10│      2│   9295│    0.18│   95.85│    4.21│54893.06│  10,362│     941│200:2;429:9292;500:3
      12s│     10│      2│  10083│    0.17│   95.85│    4.21│59903.03│  10,362│     863│200:2;429:10080;500:3
      13s│     10│      2│  10889│    0.15│   95.85│    4.21│64892.42│  10,362│     797│200:2;429:10886;500:3
      14s│     10│      2│  11383│    0.14│   95.85│    4.21│69829.26│  10,362│     740│200:2;429:11380;500:3

### 3.3.6header\_value\_match ###

{
      "descriptor_value": "...",
      "expect_match": "{...}",
      "headers": []
    }

部署ratelimit

1创建cm

cat << EOF > ratelimit-config.yaml
    apiVersion: v1
    kind: ConfigMap
    metadata:
      name: ratelimit-config
    data:
      config.yaml: |
        domain: productpage-ratelimit
        descriptors:
          - key: header_match
            value: "test"
            rate_limit:
              unit: minute
              requests_per_unit: 3
          - key: header_match
            rate_limit:
              unit: minute
              requests_per_unit: 10
    EOF
    
    kubectl apply -f ratelimit-config.yaml -n istio

2创建限速服务deployment

3创建envoy-filter

4创建action envoyfilter

cat << EOF > envoyfilter-action.yaml 
    apiVersion: networking.istio.io/v1alpha3
    kind: EnvoyFilter
    metadata:
      name: filter-ratelimit-svc
      namespace: istio
    spec:
      workloadSelector:
        labels:
          app: productpage
      configPatches:
        - applyTo: VIRTUAL_HOST
          match:
            context: SIDECAR_INBOUND
            routeConfiguration:
              vhost:
                name: "inbound|http|9080"
                route:
                  action: ANY
          patch:
            operation: MERGE
            # Applies the rate limit rules.
            value:
              rate_limits:
                - actions: 
                  - header_value_match:
                      descriptor_value: test
                      expect_match: true
                      headers:
                      - name: test
                        exact_match: test
    EOF
    
    kubectl apply -f envoyfilter-action.yaml  -n istio

[root@node01 ~]# go-stress-testing -n 1000000 -c 10 -u http://192.168.229.135:32688/productpage  -H "test:test"
    
     开始启动  并发数:10 请求数:1000000 请求参数: 
    request:
     form:http 
     url:http://192.168.229.135:32688/productpage 
     method:GET 
     headers:map[test:test] 
     data: 
     verify:statusCode 
     timeout:30s 
     debug:false 
    
    
    
    ─────┬───────┬───────┬───────┬────────┬────────┬────────┬────────┬────────┬────────┬────────
     耗时│ 并发数│ 成功数│ 失败数│   qps  │最长耗时│最短耗时│平均耗时│下载字节│字节每秒│ 错误码
    ─────┼───────┼───────┼───────┼────────┼────────┼────────┼────────┼────────┼────────┼────────
       1s│     10│      0│    551│    0.00│   46.14│    4.79│    0.00│        │        │429:551
       2s│     10│      0│   1058│    0.00│   46.14│    4.79│    0.00│        │        │429:1058
       3s│     10│      3│   1505│    1.01│   93.00│    4.79│ 9918.12│  14,545│   4,847│200:3;429:1505
       4s│     10│      3│   1996│    0.75│   93.00│    4.79│13286.69│  14,545│   3,635│200:3;429:1996
       5s│     10│      3│   2491│    0.60│   93.00│    4.79│16616.88│  14,545│   2,908│200:3;429:2491
       6s│     10│      3│   2962│    0.50│   93.00│    4.79│19942.64│  14,545│   2,423│200:3;429:2962
       7s│     10│      3│   3481│    0.43│   93.00│    4.79│23293.64│  14,545│   2,077│200:3;429:3481

### 3.3.7dynamic\_metadata ###

已废弃

### 3.3.8metadata ###

{
      "descriptor_key": "...",
      "metadata_key": "{...}",
      "default_value": "...",
      "source": "..."
    }

source：

*  DYNAMIC
    
    *(DEFAULT)* ⁣Query [dynamic metadata][]
 *  ROUTE\_ENTRY
    
    ⁣Query [route entry metadata][]

部署ratelimit

1创建cm

cat << EOF > ratelimit-config.yaml
    apiVersion: v1
    kind: ConfigMap
    metadata:
      name: ratelimit-config
    data:
      config.yaml: |
        domain: productpage-ratelimit
        descriptors:
          - key: test
            value: "test"
            rate_limit:
              unit: minute
              requests_per_unit: 4
          - key: test
            rate_limit:
              unit: minute
              requests_per_unit: 10
    EOF
    
    kubectl apply -f ratelimit-config.yaml -n istio

2创建限速服务deployment

3创建envoy-filter

4创建action envoyfilter

cat << EOF > envoyfilter-action.yaml 
    apiVersion: networking.istio.io/v1alpha3
    kind: EnvoyFilter
    metadata:
      name: filter-ratelimit-svc
      namespace: istio
    spec:
      workloadSelector:
        labels:
          app: productpage
      configPatches:
        - applyTo: VIRTUAL_HOST
          match:
            context: SIDECAR_INBOUND
            routeConfiguration:
              vhost:
                name: "inbound|http|9080"
                route:
                  action: ANY
          patch:
            operation: MERGE
            # Applies the rate limit rules.
            value:
              rate_limits:
                - actions: 
                  - metadata:
                      descriptor_key: test
                      metadata_key:
                        key: envoy.xxx
                        path:
                        - key: prop
                        - key: foo
                      default_value: test
                      source: DYNAMIC
    EOF
    
    kubectl apply -f envoyfilter-action.yaml  -n istio

### 3.3.9extension ###

extensions.rate\_limit\_descriptors.expr.v3.Descriptor

{
      "descriptor_key": "...",
      "skip_if_error": "...",
      "text": "...",
      "parsed": "{...}"
    }

部署ratelimit

1创建cm

cat << EOF > ratelimit-config.yaml
    apiVersion: v1
    kind: ConfigMap
    metadata:
      name: ratelimit-config
    data:
      config.yaml: |
        domain: productpage-ratelimit
        descriptors:
          - key: test
            value: "outbound_.9080_._.productpage.istio.svc.cluster.local"
            rate_limit:
              unit: minute
              requests_per_unit: 4
          - key: test
            rate_limit:
              unit: minute
              requests_per_unit: 10
    EOF
    
    kubectl apply -f ratelimit-config.yaml -n istio

2创建限速服务deployment

3创建envoy-filter

4创建action envoyfilter

cat << EOF > envoyfilter-action.yaml 
    apiVersion: networking.istio.io/v1alpha3
    kind: EnvoyFilter
    metadata:
      name: filter-ratelimit-svc
      namespace: istio
    spec:
      workloadSelector:
        labels:
          app: productpage
      configPatches:
        - applyTo: VIRTUAL_HOST
          match:
            context: SIDECAR_INBOUND
            routeConfiguration:
              vhost:
                name: "inbound|http|9080"
                route:
                  action: ANY
          patch:
            operation: MERGE
            # Applies the rate limit rules.
            value:
              rate_limits:
                - actions: 
                  - extension:
                      name: envoy.rate_limit_descriptors.expr
                      typed_config:
                        "@type": type.googleapis.com/envoy.extensions.rate_limit_descriptors.expr.v3.Descriptor
                        descriptor_key: test
                        skip_if_error: true
                        text: "connection.requested_server_name"
    EOF
    
    kubectl apply -f envoyfilter-action.yaml  -n istio

[root@node01 ~]# go-stress-testing -n 1000000 -c 10 -u http://192.168.229.135:32688/productpage
    
     开始启动  并发数:10 请求数:1000000 请求参数: 
    request:
     form:http 
     url:http://192.168.229.135:32688/productpage 
     method:GET 
     headers:map[] 
     data: 
     verify:statusCode 
     timeout:30s 
     debug:false 
    
    
    
    ─────┬───────┬───────┬───────┬────────┬────────┬────────┬────────┬────────┬────────┬────────
     耗时│ 并发数│ 成功数│ 失败数│   qps  │最长耗时│最短耗时│平均耗时│下载字节│字节每秒│ 错误码
    ─────┼───────┼───────┼───────┼────────┼────────┼────────┼────────┼────────┼────────┼────────
       1s│     10│      0│    850│    0.00│   28.69│    3.87│    0.00│        │        │429:850
       2s│     10│      0│   1688│    0.00│   32.26│    3.87│    0.00│        │        │429:1688
       3s│     10│      0│   2593│    0.00│   32.26│    3.87│    0.00│        │        │429:2593
       4s│     10│      4│   3443│    1.00│   97.18│    3.87│ 9975.97│  19,724│   4,930│200:4;429:3440;500:3
       5s│     10│      4│   4005│    0.80│   97.18│    3.87│12431.83│  19,724│   3,944│200:4;429:4002;500:3
       6s│     10│      4│   4464│    0.67│   97.18│    3.87│14930.66│  19,724│   3,287│200:4;429:4461;500:3
       7s│     10│      4│   4962│    0.57│   97.18│    3.87│17428.69│  19,724│   2,817│200:4;429:4959;500:3
       8s│     10│      4│   5477│    0.50│   97.18│    3.87│19941.18│  19,724│   2,464│200:4;429:5474;500:3

## 3.4按请求类型限流 ##

### 3.4.1internal ###

部署ratelimit

1创建cm

cat << EOF > ratelimit-config.yaml
    apiVersion: v1
    kind: ConfigMap
    metadata:
      name: ratelimit-config
    data:
      config.yaml: |
        domain: productpage-ratelimit
        descriptors:
          - key: header_match
            value: "test"
            rate_limit:
              unit: minute
              requests_per_unit: 3
          - key: header_match
            rate_limit:
              unit: minute
              requests_per_unit: 10
    EOF
    
    kubectl apply -f ratelimit-config.yaml -n istio

2创建限速服务deployment

3创建envoy-filter

cat << EOF > envoyfilter-filter.yaml
    apiVersion: networking.istio.io/v1alpha3
    kind: EnvoyFilter
    metadata:
      name: filter-ratelimit
      namespace: istio
    spec:
      workloadSelector:
        # select by label in the same namespace
        labels:
          app: productpage
      configPatches:
        # The Envoy config you want to modify
        - applyTo: HTTP_FILTER
          match:
            context: SIDECAR_INBOUND
            listener:
              filterChain:
                filter:
                  name: "envoy.filters.network.http_connection_manager"
                  subFilter:
                    name: "envoy.filters.http.router"
          patch:
            operation: INSERT_BEFORE
            # Adds the Envoy Rate Limit Filter in HTTP filter chain.
            value:
              name: envoy.filters.http.ratelimit
              typed_config:
                "@type": type.googleapis.com/envoy.extensions.filters.http.ratelimit.v3.RateLimit
                # domain can be anything! Match it to the ratelimter service config
                domain: productpage-ratelimit
                failure_mode_deny: true
                request_type: internal
                timeout: 20ms
                rate_limited_as_resource_exhausted: true
                enable_x_ratelimit_headers: DRAFT_VERSION_03
                disable_x_envoy_ratelimited_header: false
                rate_limit_service:
                  grpc_service:
                    envoy_grpc:
                      cluster_name: rate_limit_cluster
                    timeout: 10s
                  transport_api_version: V3
        - applyTo: CLUSTER
          match:
            cluster:
              service: ratelimit.istio.svc.cluster.local
          patch:
            operation: ADD
            # Adds the rate limit service cluster for rate limit service defined in step 1.
            value:
              name: rate_limit_cluster
              type: STRICT_DNS
              connect_timeout: 10s
              lb_policy: ROUND_ROBIN
              http2_protocol_options: {}
              load_assignment:
                cluster_name: rate_limit_cluster
                endpoints:
                - lb_endpoints:
                  - endpoint:
                      address:
                         socket_address:
                          address: ratelimit.istio.svc.cluster.local
                          port_value: 8081
    EOF
    
    kubectl apply -f envoyfilter-filter.yaml -n istio

4创建action envoyfilter

cat << EOF > envoyfilter-action.yaml 
    apiVersion: networking.istio.io/v1alpha3
    kind: EnvoyFilter
    metadata:
      name: filter-ratelimit-svc
      namespace: istio
    spec:
      workloadSelector:
        labels:
          app: productpage
      configPatches:
        - applyTo: VIRTUAL_HOST
          match:
            context: SIDECAR_INBOUND
            routeConfiguration:
              vhost:
                name: "inbound|http|9080"
                route:
                  action: ANY
          patch:
            operation: MERGE
            # Applies the rate limit rules.
            value:
              rate_limits:
                - actions: 
                  - header_value_match:
                      descriptor_value: test
                      expect_match: true
                      headers:
                      - name: test
                        exact_match: test
    EOF
    
    kubectl apply -f envoyfilter-action.yaml  -n istio

[root@node01 ~]# go-stress-testing -n 1000000 -c 10 -u http://192.168.229.135:32688/productpage -H "test:test"
    
     开始启动  并发数:10 请求数:1000000 请求参数: 
    request:
     form:http 
     url:http://192.168.229.135:32688/productpage 
     method:GET 
     headers:map[test:test] 
     data: 
     verify:statusCode 
     timeout:30s 
     debug:false 
    
    
    
    ─────┬───────┬───────┬───────┬────────┬────────┬────────┬────────┬────────┬────────┬────────
     耗时│ 并发数│ 成功数│ 失败数│   qps  │最长耗时│最短耗时│平均耗时│下载字节│字节每秒│ 错误码
    ─────┼───────┼───────┼───────┼────────┼────────┼────────┼────────┼────────┼────────┼────────
       1s│     10│      0│    402│    0.00│   58.08│    8.42│    0.00│        │        │429:402
       2s│     10│      0│    838│    0.00│   58.21│    7.16│    0.00│        │        │429:838
       3s│     10│      0│   1345│    0.00│   58.21│    7.16│    0.00│        │        │429:1345
       4s│     10│      0│   1813│    0.00│   58.21│    6.42│    0.00│        │        │429:1813
       5s│     10│      3│   2284│    0.60│   94.15│    6.42│16583.06│  14,545│   2,908│200:3;429:2284
       6s│     10│      3│   2750│    0.50│   94.15│    6.42│19914.90│  14,545│   2,424│200:3;429:2750
       7s│     10│      3│   3189│    0.43│   94.15│    6.42│23246.60│  14,545│   2,077│200:3;429:3189
       8s│     10│      3│   3616│    0.38│   94.15│    6.42│26548.82│  14,545│   1,817│200:3;429:3616

### 3.4.2external ###

部署ratelimit

1创建cm

cat << EOF > ratelimit-config.yaml
    apiVersion: v1
    kind: ConfigMap
    metadata:
      name: ratelimit-config
    data:
      config.yaml: |
        domain: productpage-ratelimit
        descriptors:
          - key: header_match
            value: "test"
            rate_limit:
              unit: minute
              requests_per_unit: 3
          - key: header_match
            rate_limit:
              unit: minute
              requests_per_unit: 10
    EOF
    
    kubectl apply -f ratelimit-config.yaml -n istio

2创建限速服务deployment

3创建envoy-filter

cat << EOF > envoyfilter-filter.yaml
    apiVersion: networking.istio.io/v1alpha3
    kind: EnvoyFilter
    metadata:
      name: filter-ratelimit
      namespace: istio
    spec:
      workloadSelector:
        # select by label in the same namespace
        labels:
          app: productpage
      configPatches:
        # The Envoy config you want to modify
        - applyTo: HTTP_FILTER
          match:
            context: SIDECAR_INBOUND
            listener:
              filterChain:
                filter:
                  name: "envoy.filters.network.http_connection_manager"
                  subFilter:
                    name: "envoy.filters.http.router"
          patch:
            operation: INSERT_BEFORE
            # Adds the Envoy Rate Limit Filter in HTTP filter chain.
            value:
              name: envoy.filters.http.ratelimit
              typed_config:
                "@type": type.googleapis.com/envoy.extensions.filters.http.ratelimit.v3.RateLimit
                # domain can be anything! Match it to the ratelimter service config
                domain: productpage-ratelimit
                failure_mode_deny: true
                request_type: external
                timeout: 20ms
                rate_limited_as_resource_exhausted: true
                enable_x_ratelimit_headers: DRAFT_VERSION_03
                disable_x_envoy_ratelimited_header: false
                rate_limit_service:
                  grpc_service:
                    envoy_grpc:
                      cluster_name: rate_limit_cluster
                    timeout: 10s
                  transport_api_version: V3
        - applyTo: CLUSTER
          match:
            cluster:
              service: ratelimit.istio.svc.cluster.local
          patch:
            operation: ADD
            # Adds the rate limit service cluster for rate limit service defined in step 1.
            value:
              name: rate_limit_cluster
              type: STRICT_DNS
              connect_timeout: 10s
              lb_policy: ROUND_ROBIN
              http2_protocol_options: {}
              load_assignment:
                cluster_name: rate_limit_cluster
                endpoints:
                - lb_endpoints:
                  - endpoint:
                      address:
                         socket_address:
                          address: ratelimit.istio.svc.cluster.local
                          port_value: 8081
    EOF
    
    kubectl apply -f envoyfilter-filter.yaml -n istio

4创建action envoyfilter

cat << EOF > envoyfilter-action.yaml 
    apiVersion: networking.istio.io/v1alpha3
    kind: EnvoyFilter
    metadata:
      name: filter-ratelimit-svc
      namespace: istio
    spec:
      workloadSelector:
        labels:
          app: productpage
      configPatches:
        - applyTo: VIRTUAL_HOST
          match:
            context: SIDECAR_INBOUND
            routeConfiguration:
              vhost:
                name: "inbound|http|9080"
                route:
                  action: ANY
          patch:
            operation: MERGE
            # Applies the rate limit rules.
            value:
              rate_limits:
                - actions: 
                  - header_value_match:
                      descriptor_value: test
                      expect_match: true
                      headers:
                      - name: test
                        exact_match: test
    EOF
    
    kubectl apply -f envoyfilter-action.yaml  -n istio

[root@node01 ~]# go-stress-testing -n 1000000 -c 10 -u http://192.168.229.135:32688/productpage -H "test:test"
    
     开始启动  并发数:10 请求数:1000000 请求参数: 
    request:
     form:http 
     url:http://192.168.229.135:32688/productpage 
     method:GET 
     headers:map[test:test] 
     data: 
     verify:statusCode 
     timeout:30s 
     debug:false 
    
    
    
    ─────┬───────┬───────┬───────┬────────┬────────┬────────┬────────┬────────┬────────┬────────
     耗时│ 并发数│ 成功数│ 失败数│   qps  │最长耗时│最短耗时│平均耗时│下载字节│字节每秒│ 错误码
    ─────┼───────┼───────┼───────┼────────┼────────┼────────┼────────┼────────┼────────┼────────
       1s│     10│     94│      0│   98.74│  150.75│   32.92│  101.27│ 455,078│ 454,974│200:94
       2s│     10│    191│      0│   97.52│  189.63│   32.92│  102.55│ 925,701│ 462,829│200:191
       3s│     10│    288│      0│   98.58│  189.63│   32.92│  101.44│1,396,320│ 465,432│200:288
       4s│     10│    390│      0│   99.04│  189.63│   32.92│  100.97│1,890,850│ 472,669│200:390

因为请求是内部的，所以这里没有ratelimit

### 3.4.3both ###

部署ratelimit

1创建cm

cat << EOF > ratelimit-config.yaml
    apiVersion: v1
    kind: ConfigMap
    metadata:
      name: ratelimit-config
    data:
      config.yaml: |
        domain: productpage-ratelimit
        descriptors:
          - key: header_match
            value: "test"
            rate_limit:
              unit: minute
              requests_per_unit: 3
          - key: header_match
            rate_limit:
              unit: minute
              requests_per_unit: 10
    EOF
    
    kubectl apply -f ratelimit-config.yaml -n istio

2创建限速服务deployment

3创建envoy-filter

cat << EOF > envoyfilter-filter.yaml
    apiVersion: networking.istio.io/v1alpha3
    kind: EnvoyFilter
    metadata:
      name: filter-ratelimit
      namespace: istio
    spec:
      workloadSelector:
        # select by label in the same namespace
        labels:
          app: productpage
      configPatches:
        # The Envoy config you want to modify
        - applyTo: HTTP_FILTER
          match:
            context: SIDECAR_INBOUND
            listener:
              filterChain:
                filter:
                  name: "envoy.filters.network.http_connection_manager"
                  subFilter:
                    name: "envoy.filters.http.router"
          patch:
            operation: INSERT_BEFORE
            # Adds the Envoy Rate Limit Filter in HTTP filter chain.
            value:
              name: envoy.filters.http.ratelimit
              typed_config:
                "@type": type.googleapis.com/envoy.extensions.filters.http.ratelimit.v3.RateLimit
                # domain can be anything! Match it to the ratelimter service config
                domain: productpage-ratelimit
                failure_mode_deny: true
                request_type: both
                timeout: 20ms
                rate_limited_as_resource_exhausted: true
                enable_x_ratelimit_headers: DRAFT_VERSION_03
                disable_x_envoy_ratelimited_header: false
                rate_limit_service:
                  grpc_service:
                    envoy_grpc:
                      cluster_name: rate_limit_cluster
                    timeout: 10s
                  transport_api_version: V3
        - applyTo: CLUSTER
          match:
            cluster:
              service: ratelimit.istio.svc.cluster.local
          patch:
            operation: ADD
            # Adds the rate limit service cluster for rate limit service defined in step 1.
            value:
              name: rate_limit_cluster
              type: STRICT_DNS
              connect_timeout: 10s
              lb_policy: ROUND_ROBIN
              http2_protocol_options: {}
              load_assignment:
                cluster_name: rate_limit_cluster
                endpoints:
                - lb_endpoints:
                  - endpoint:
                      address:
                         socket_address:
                          address: ratelimit.istio.svc.cluster.local
                          port_value: 8081
    EOF
    
    kubectl apply -f envoyfilter-filter.yaml -n istio

4创建action envoyfilter

cat << EOF > envoyfilter-action.yaml 
    apiVersion: networking.istio.io/v1alpha3
    kind: EnvoyFilter
    metadata:
      name: filter-ratelimit-svc
      namespace: istio
    spec:
      workloadSelector:
        labels:
          app: productpage
      configPatches:
        - applyTo: VIRTUAL_HOST
          match:
            context: SIDECAR_INBOUND
            routeConfiguration:
              vhost:
                name: "inbound|http|9080"
                route:
                  action: ANY
          patch:
            operation: MERGE
            # Applies the rate limit rules.
            value:
              rate_limits:
                - actions: 
                  - header_value_match:
                      descriptor_value: test
                      expect_match: true
                      headers:
                      - name: test
                        exact_match: test
    EOF
    
    kubectl apply -f envoyfilter-action.yaml  -n istio

[root@node01 ~]# go-stress-testing -n 1000000 -c 10 -u http://192.168.229.135:32688/productpage -H "test:test"
    
     开始启动  并发数:10 请求数:1000000 请求参数: 
    request:
     form:http 
     url:http://192.168.229.135:32688/productpage 
     method:GET 
     headers:map[test:test] 
     data: 
     verify:statusCode 
     timeout:30s 
     debug:false 
    
    
    
    ─────┬───────┬───────┬───────┬────────┬────────┬────────┬────────┬────────┬────────┬────────
     耗时│ 并发数│ 成功数│ 失败数│   qps  │最长耗时│最短耗时│平均耗时│下载字节│字节每秒│ 错误码
    ─────┼───────┼───────┼───────┼────────┼────────┼────────┼────────┼────────┼────────┼────────
       1s│     10│      0│    832│    0.00│   30.45│    4.18│    0.00│        │        │429:832
       2s│     10│      0│   1621│    0.00│   30.45│    4.18│    0.00│        │        │429:1621
       3s│     10│      0│   2398│    0.00│   30.45│    4.18│    0.00│        │        │429:2398
       4s│     10│      0│   3188│    0.00│   30.45│    3.86│    0.00│        │        │429:3179;500:9
       5s│     10│      0│   4087│    0.00│   30.45│    3.86│    0.00│        │        │429:4078;500:9
       6s│     10│      3│   4961│    0.50│   87.15│    3.86│19948.44│  14,545│   2,424│200:3;429:4952;500:9
       7s│     10│      3│   5843│    0.43│   87.15│    3.86│23253.71│  14,545│   2,077│200:3;429:5834;500:9
       8s│     10│      3│   6673│    0.38│   87.15│    3.86│26606.37│  14,545│   1,818│200:3;429:6664;500:9

## 3.5stage ##

部署ratelimit

1创建cm

cat << EOF > ratelimit-config.yaml
    apiVersion: v1
    kind: ConfigMap
    metadata:
      name: ratelimit-config
    data:
      config.yaml: |
        domain: productpage-ratelimit
        descriptors:
          - key: remote_address
            value: "172.20.1.0"
            rate_limit:
              unit: minute
              requests_per_unit: 1
          - key: remote_address
            rate_limit:
              unit: minute
              requests_per_unit: 10
    EOF
    
    kubectl apply -f ratelimit-config.yaml -n istio

2创建限速服务deployment

3创建envoy-filter

cat << EOF > envoyfilter-filter.yaml
    apiVersion: networking.istio.io/v1alpha3
    kind: EnvoyFilter
    metadata:
      name: filter-ratelimit
      namespace: istio
    spec:
      workloadSelector:
        # select by label in the same namespace
        labels:
          app: productpage
      configPatches:
        # The Envoy config you want to modify
        - applyTo: HTTP_FILTER
          match:
            context: SIDECAR_INBOUND
            listener:
              filterChain:
                filter:
                  name: "envoy.filters.network.http_connection_manager"
                  subFilter:
                    name: "envoy.filters.http.router"
          patch:
            operation: INSERT_BEFORE
            # Adds the Envoy Rate Limit Filter in HTTP filter chain.
            value:
              name: envoy.filters.http.ratelimit
              typed_config:
                "@type": type.googleapis.com/envoy.extensions.filters.http.ratelimit.v3.RateLimit
                # domain can be anything! Match it to the ratelimter service config
                domain: productpage-ratelimit
                stage: 1
                failure_mode_deny: true
                rate_limit_service:
                  grpc_service:
                    envoy_grpc:
                      cluster_name: rate_limit_cluster
                    timeout: 10s
                  transport_api_version: V3
        - applyTo: CLUSTER
          match:
            cluster:
              service: ratelimit.istio.svc.cluster.local
          patch:
            operation: ADD
            # Adds the rate limit service cluster for rate limit service defined in step 1.
            value:
              name: rate_limit_cluster
              type: STRICT_DNS
              connect_timeout: 10s
              lb_policy: ROUND_ROBIN
              http2_protocol_options: {}
              load_assignment:
                cluster_name: rate_limit_cluster
                endpoints:
                - lb_endpoints:
                  - endpoint:
                      address:
                         socket_address:
                          address: ratelimit.istio.svc.cluster.local
                          port_value: 8081
    EOF
    
    kubectl apply -f envoyfilter-filter.yaml -n istio

4创建action envoyfilter

cat << EOF > envoyfilter-action.yaml 
    apiVersion: networking.istio.io/v1alpha3
    kind: EnvoyFilter
    metadata:
      name: filter-ratelimit-svc
      namespace: istio
    spec:
      workloadSelector:
        labels:
          app: productpage
      configPatches:
        - applyTo: VIRTUAL_HOST
          match:
            context: SIDECAR_INBOUND
            routeConfiguration:
              vhost:
                name: "inbound|http|9080"
                route:
                  action: ANY
          patch:
            operation: MERGE
            # Applies the rate limit rules.
            value:
              rate_limits:
                - actions: 
                  - remote_address: {}
                  stage: 1
                  disable_key: test
                  limit:
                    dynamic_metadata: 
                      metadata_key:
                        key: envoy.xxx
                        path:
                        - key: prop
                        - key: foo
    EOF
    
    kubectl apply -f envoyfilter-action.yaml  -n istio

[root@node01 ~]# go-stress-testing -n 1000000 -c 10 -u http://192.168.229.135:32688/productpage
    
     开始启动  并发数:10 请求数:1000000 请求参数: 
    request:
     form:http 
     url:http://192.168.229.135:32688/productpage 
     method:GET 
     headers:map[] 
     data: 
     verify:statusCode 
     timeout:30s 
     debug:false 
    
    
    
    ─────┬───────┬───────┬───────┬────────┬────────┬────────┬────────┬────────┬────────┬────────
     耗时│ 并发数│ 成功数│ 失败数│   qps  │最长耗时│最短耗时│平均耗时│下载字节│字节每秒│ 错误码
    ─────┼───────┼───────┼───────┼────────┼────────┼────────┼────────┼────────┼────────┼────────
       1s│     10│      1│    835│    1.01│   44.86│    5.45│ 9941.10│   4,183│   4,182│200:1;429:815;500:20
       2s│     10│      1│   1711│    0.50│   44.86│    4.96│19839.11│   4,183│   2,091│200:1;429:1688;500:23
       3s│     10│      1│   2465│    0.33│   44.86│    4.96│29876.69│   4,183│   1,393│200:1;429:2442;500:23
       4s│     10│      1│   3007│    0.25│   66.38│    4.96│39852.79│   4,183│   1,045│200:1;429:2984;500:23
       5s│     10│      1│   3852│    0.20│   66.38│    4.55│49888.50│   4,183│     836│200:1;429:3829;500:23
       6s│     10│      1│   4733│    0.17│   66.38│    4.55│59809.06│   4,183│     696│200:1;429:4710;500:23
       7s│     10│      1│   5581│    0.14│   66.38│    4.55│69770.80│   4,183│     597│200:1;429:5558;500:23
       8s│     10│      1│   6497│    0.13│   66.38│    4.55│79826.00│   4,183│     522│200:1;429:6474;500:23
       9s│     10│      1│   7407│    0.11│   66.38│    4.55│89840.05│   4,183│     464│200:1;429:7384;500:23
      10s│     10│      1│   8271│    0.10│   66.38│    4.55│99801.86│   4,183│     418│200:1;429:8248;500:23

# 4清理 #

[watermark_type_ZHJvaWRzYW5zZmFsbGJhY2s_shadow_50_text_Q1NETiBAaHhwamF2YTE_size_8_color_FFFFFF_t_70_g_se_x_16]: /images/20220828/e303e86bfe2f46948138eb59057324cf.png
[istio_50]: https://mp.weixin.qq.com/s/_OPMjmWCDSTlmbAxytg33A
[istio]: https://mp.weixin.qq.com/s/kjvtno_UzJAGr1joxm3tvw
[istio_istio]: https://mp.weixin.qq.com/s/e2-p_NY31Mc5y07S4Bb1JQ
[istio 1]: https://mp.weixin.qq.com/s/iTV0SCQLhSOTYncp3PxoHQ
[istio 2]: https://mp.weixin.qq.com/s/eINlrT242RlAwfuZu1_PcA
[envoyfilter_istio_-http-rbac-_AuthorizationPolicy]: https://mp.weixin.qq.com/s/Q-yTGBNqkjOov49LTXPrPA
[envoyfilter_istio_-02-http-corsFilter-_vs]: https://mp.weixin.qq.com/s/bFMdrT7OKSqoWv8pDXOIqQ
[envoyfilter_istio_-03-http-csrf filter-_csrf]: https://mp.weixin.qq.com/s/dlG9phV-kMMxs4UPzuRNYg
[envoyfilter_istio_http-jwt_authn-_RequestAuthorization]: https://mp.weixin.qq.com/s/IfrJ-FdLYsdhPkDrcO_YXA
[envoyfilter_istio_-05-fault-filter-_vs]: https://mp.weixin.qq.com/s/a4gDXqD8uswgglMwkUph2Q
[envoyfilter_istio_-06-http-match-_vs]: https://mp.weixin.qq.com/s/TpS3DHdcaJiPivIS_jsIGw
[envoyfilter_istio_-07-_dr]: https://mp.weixin.qq.com/s/k-zJm7n7OOKg6MotQJYPAA
[envoyfilter_istio_-08-]: https://mp.weixin.qq.com/s/n_bJ3saNeTtXRprIymhwzg
[envoyfilter_istio_-09-http-route filter]: https://mp.weixin.qq.com/s/JAeZr0BMVQxIrB3S_7Ru0A
[envoyfilter_istio_-network filter-redis proxy]: https://mp.weixin.qq.com/s/B9IPTI8F3q42nVkqbQRlPw
[envoyfilter_istio_-network filter-HttpConnectionManager]: https://mp.weixin.qq.com/s/W3OPQWkCVvuHGS49cTNhlg
[envoyfilter_istio_-ratelimit-istio ratelimit]: https://mp.weixin.qq.com/s/KgiDsELNol27mgZvDbFsqw
[draft RFC Version 03]: https://tools.ietf.org/id/draft-polli-ratelimit-headers-03.html
[watermark_type_ZHJvaWRzYW5zZmFsbGJhY2s_shadow_50_text_Q1NETiBAaHhwamF2YTE_size_20_color_FFFFFF_t_70_g_se_x_16]: /images/20220828/22a9bc0a3ee942abb57280bd90b9ee38.png
[watermark_type_ZHJvaWRzYW5zZmFsbGJhY2s_shadow_50_text_Q1NETiBAaHhwamF2YTE_size_20_color_FFFFFF_t_70_g_se_x_16 1]: /images/20220828/a77e313f006844938711d182a7b66502.png
[watermark_type_ZHJvaWRzYW5zZmFsbGJhY2s_shadow_50_text_Q1NETiBAaHhwamF2YTE_size_20_color_FFFFFF_t_70_g_se_x_16 2]: /images/20220828/c7b3266f49174a60b239a43f5fbaed66.png
[watermark_type_ZHJvaWRzYW5zZmFsbGJhY2s_shadow_50_text_Q1NETiBAaHhwamF2YTE_size_20_color_FFFFFF_t_70_g_se_x_16 3]: /images/20220828/44cd30db70404ac4a98565fbcece95d0.png
[dynamic metadata]: https://www.envoyproxy.io/docs/envoy/latest/configuration/advanced/well_known_dynamic_metadata#well-known-dynamic-metadata
[route entry metadata]: https://www.envoyproxy.io/docs/envoy/latest/api-v3/config/route/v3/route_components.proto.html?highlight=rate_limits#envoy-v3-api-field-config-route-v3-route-metadata