Envoy常用术语
envoy文档官网 Life of a Request — envoy 1.31.0-dev-e543e1 documentationhttps://www.envoyproxy.io/docs/envoy/latest/intro/life_of_a_request#terminology
基础总结
(1)Envoy
Envoy自己本身是工作在L7层的一个proxy(官方也是这样说明的),在7层它支持HTTP、HTTPS,还有Mongo、Mysql等filter。
但是,要想7层正常工作,必须要在L3或L4层,也就是网络层或传输层,去激活一些filter才行,比如常用的HTTP Connection Manager。
虽然看名称像是7层的HTTP Connection Manager,但实际上它是工作在4层的,将内核本身就能够处理的传输层以下的报文,再一次向上提升到应用层。(2)调用filter
必须要指明name,并且要注意,各个filter.name,都有其固定的名称,也就是这个name是不能随便定义和修改的,如下:- filters:- name: envoy.filters.network.http_connection_manager并且,各filter还需要适配不同的API版本,这个API版本也是固定的,不能随意更改(不同API版本语法参数不同),如下:typed_config: # 下面这个@type,也是一个固定且必须使用的key,用来指明我们要使用的API版本"@type": type.googleapis.com/envoy.extensions.filters.network.http_connection_manager.v3.HttpConnectionManager # 这是这个 v3.HttpConnectionManager ,现在基本都是V3,V1V2已经停止维护了每个filter还可以引入其专用的配置参数,那这些配置参数就能形成一些新的、可以插入到我们配置文件当中的配置段,那些可以通过xDS API所发现的配置,通常不止一段配置的时候,每个配置段都可以已通过一个名称经进行标识,这个名称就没有什么固定的要求。如下:listeners:- name: listener_0route_config:name: local_routeclusters:- name: web_cluster
工作逻辑
Envoy作为一个7层的反向代理来说,就相当于一个中介,连接着下游客户端(DownStream)和上游服务端(UPStream)。
Envoy面向下游客户端的时候,它必须要监听一个套接字(Socket),如下:listeners:- name: listener_0address:socket_address: { address: 127.0.0.1, port_value: 80 }Envoy需要通过一个listener来接入流量,而对listener而言(7层 HTTP代理),内部可以配置一到多个虚拟主机(virtual_hosts),凡是通过listener进来的请求,都会通过请求报文中的host,来获取请求的目标主机。:virtual_hosts:- name: web_service_1domains: ["*.ik8s.io", "ik8s.io"] 而一个virtual_hosts内部,最为重要的一个配置,就是与该virtual_hosts相关的路由(route),只有配置了路由,Envoy才知道把请求代理到后端的哪一个集群中,如下:routes:- match: { prefix: "/" }route: { cluster: web_cluster }对于下游来说,listeners这里最重要的几个配置,就是如下几个:
listeners.name.address.socket_address
filter_chains.filters # 注意:代理型的过滤器,在一个listeners中只需要配置一个对于上游来说,重要的配置就是如下几个:
clusters.name # 集群名称
clusters.lb_policy # 负载均衡策略
clusters.load_assignment # 端点配置
clusters.type # 端点获取放回寺。有几种获取端点的配置如下:
(1)STATIC:手动配置。
(2)STRICT_DNS:会尝试把域名解析后的每一个IP地址,都配置为一个端点(适合局域网使用)。
(3)LOGICAL_DNS:和STRICT_DNS不同,只会解析后的第一个IP地址,配置为端点(适合跨互联网使用)。
(4)EDS:这是envoy原生的端点发现方式。
代理总结
Envoy部署分为3种类型:
(1)前端代理,也可成为边缘代理(Front/Edge Proxy)
这个时候它扮演的角色其实就是网关,但是网关也分为入向流量网关和出向流量网关。(2)服务到服务(Service to Service Only)
这个时候的Envoy代理被称为:网格内部的数据平面(网格代理),这个时候每一个实例(服务),都应该有一个专属的代理,就是前面说的边车模式的pod。
如果按照侦听器划分,那就是:
Ingress Listener:入向流量侦听器,负责反向代理
Egress Listener:出向流量侦听器,负责正向代理(3)Egress Listener To External:出向流量正向代理到集群外部服务
这个是可选的,不一定必须得要。网关分类,就是说边缘代理,也可以分为2种类型:
(1)Ingress Gateway:入向流量网关(2)Egress Gateway:出向流量网关
当我们不需要对出向流量做什么特殊处理的时候,其实只需要有入向流量网关就够了。
如果需要对出向流量做治理的时候,才需要出向流量网关
docker 镜像拉取
# docker pull envoyproxy/envoy:latest
latest: Pulling from envoyproxy/envoy
fe703b657a32: Pull complete
f9df1fafd224: Pull complete
a645a4b887f9: Pull complete
57db7fe0b522: Pull complete
8239c375d005: Pull complete
dadf1a36dc1a: Pull complete
aaa45c35c46c: Pull complete
2927ac61eca5: Pull complete
d3f2b341deed: Pull complete
Digest: sha256:ccaf9e0135bf498fb8396ad49defd7f8567bf706411d9a707efb3978fb842c89
Status: Downloaded newer image for envoyproxy/envoy:latest
docker.io/envoyproxy/envoy:latest
容器部署Envoy
1.1.1 下载资源清单
地址:https://github.com/iKubernetes/servicemesh_in_practise/tree/MageEdu_N66
1.1.2 上传包
$ ll
total 40
drwxrwxr-x 11 root root 198 May 30 14:43 Cluster-Manager
drwxrwxr-x 7 root root 126 May 30 14:43 Dynamic-Configuration
drwxrwxr-x 2 root root 90 May 30 14:43 envoy-alpine
drwxrwxr-x 9 root root 173 May 30 14:43 Envoy-Basics
drwxrwxr-x 6 root root 103 May 30 14:43 Envoy-Mesh
drwxrwxr-x 10 root root 211 May 30 14:43 HTTP-Connection-Manager
-rw-rw-r-- 1 root root 35184 May 30 14:43 LICENSE
drwxrwxr-x 11 root root 237 May 30 14:43 Monitoring-and-Tracing
-rw-rw-r-- 1 root root 2125 May 30 14:43 README.md
drwxrwxr-x 8 root root 125 May 30 14:43 Security
drwxrwxr-x 2 root root 57 May 30 14:43 template
2 tcp proxy(L4 过滤器)
TCP代理过滤器在下游客户端及上游集群之间执行1:1网络连接代理。
它可以单独用作隧道替换,也可以同其他过滤器(如MongoDB过滤器或速率限制过滤器) 结合使用;
TCP代理过滤器严格执行由全局资源管理于为每个上游集群的全局资源管理器设定的连接限制;
TCP代理过滤器检查上游集群的资源管理器是否可以在不超过该集群的最大连接数的情况下创建连接;
TCP代理过滤器可直接将请求路由至指定的集群,也能够在多个目标集群间基于权重进行调度转发;
2.1.1 cat envoy.yaml
# cat envoy.yaml static_resources:listeners:name: listener_0address:socket_address: { address: 0.0.0.0, port_value: 80 }filter_chains:- filters:- name: envoy.tcp_proxytyped_config:"@type": type.googleapis.com/envoy.extensions.filters.network.tcp_proxy.v3.TcpProxystat_prefix: tcpcluster: local_clusterclusters:- name: local_clusterconnect_timeout: 0.25stype: STATIClb_policy: ROUND_ROBINload_assignment:cluster_name: local_clusterendpoints:- lb_endpoints:- endpoint:address:socket_address: { address: 192.168.34.61, port_value: 8080 }- endpoint:address:socket_address: { address: 192.168.34.62, port_value: 8080 }
2.1.2 cat docker-compose.yaml
# cat docker-compose.yaml
version: '3.3'services:envoy:image: envoyproxy/envoy:v1.23.4volumes:- ./envoy.yaml:/etc/envoy/envoy.yamlnetworks:envoymesh:ipv4_address: 192.168.1.60aliases:- front-proxydepends_on:- webserver01- webserver02webserver01:image: ikubernetes/demoapp:v1.0environment:- PORT=8080hostname: webserver01networks:envoymesh:ipv4_address: 192.168.1.61aliases:- webserver01webserver02:image: ikubernetes/demoapp:v1.0environment:- PORT=8080hostname: webserver02networks:envoymesh:ipv4_address: 192.168.1.62aliases:- webserver02networks:envoymesh:driver: bridgeipam:config:- subnet: 192.168.1.0/24
2.1.3 拉取服务镜像
docker pull ikubernetes/demoapp:v1.0
v1.0: Pulling from ikubernetes/demoapp
c9b1b535fdd9: Pull complete
3cbce035cd7c: Pull complete
b83463f478a5: Pull complete
34b1f286d5e2: Pull complete
Digest: sha256:6698b205eb18fb0171398927f3a35fe27676c6bf5757ef57a35a4b055badf2c3
Status: Downloaded newer image for ikubernetes/demoapp:v1.0
docker.io/ikubernetes/demoapp:v1.0
2.1.4 启动容器
$ docker-compose up -d
Creating network "tcp-front-proxy_envoymesh" with driver "bridge"
Creating tcp-front-proxy_webserver02_1 ... done
Creating tcp-front-proxy_webserver01_1 ... done
Creating tcp-front-proxy_envoy_1 ... done
2.1.5 检查进程
$ docker-compose psName Command State Ports
----------------------------------------------------------------------------------
tcp-front-proxy_envoy_1 /docker-entrypoint.sh envo ... Up 10000/tcp
tcp-front-proxy_webserver01_1 /bin/sh -c python3 /usr/lo ... Up
tcp-front-proxy_webserver02_1 /bin/sh -c python3 /usr/lo ... Up
2.1.6 访问测试,轮寻访问2个服务
$ curl 192.168.1.60
iKubernetes demoapp v1.0 !! ClientIP: 192.168.1.60, ServerName: webserver02, ServerIP: 192.168.1.62!
$ curl 192.168.1.60
iKubernetes demoapp v1.0 !! ClientIP: 192.168.1.60, ServerName: webserver01, ServerIP: 192.168.1.61!
$ curl 192.168.1.60
iKubernetes demoapp v1.0 !! ClientIP: 192.168.1.60, ServerName: webserver02, ServerIP: 192.168.1.62!
$ curl 192.168.1.60
iKubernetes demoapp v1.0 !! ClientIP: 192.168.1.60, ServerName: webserver01, ServerIP: 192.168.1.61!
2.1.7 进入容器后,可直接执行envoy命令
root@e7d74ecd3159:/usr/bin# envoy --versionenvoy version: 3504d40f752eb5c20bc2883053547717bcb92fd8/1.14.1/Clean/RELEASE/BoringSSL
2.1.8 数据流程如下
2.1.9 停止
$ docker-compose down
Stopping tcp-front-proxy_envoy_1 ... done
Stopping tcp-front-proxy_webserver02_1 ... done
Stopping tcp-front-proxy_webserver01_1 ... done
Removing tcp-front-proxy_envoy_1 ... done
Removing tcp-front-proxy_webserver02_1 ... done
Removing tcp-front-proxy_webserver01_1 ... done
Removing network tcp-front-proxy_envoymesh
3 http proxy(L7 过滤器)
cd http-front-proxy/
$ ll
total 12
-rw-rw-r-- 1 root root 862 May 30 15:36 docker-compose.yaml
-rw-rw-r-- 1 root root 1581 May 30 15:35 envoy.yaml
-rw-rw-r-- 1 root root 572 May 30 14:43 README.md
3.1.1 cat envoy.yaml
$ cat envoy.yaml
static_resources:listeners:- name: listener_0address:socket_address: { address: 0.0.0.0, port_value: 80 }filter_chains:- filters:- name: envoy.filters.network.http_connection_managertyped_config:"@type": type.googleapis.com/envoy.extensions.filters.network.http_connection_manager.v3.HttpConnectionManagerstat_prefix: ingress_httpcodec_type: AUTOroute_config:name: local_routevirtual_hosts:- name: web_service_1domains: ["*.ik8s.io", "ik8s.io"]routes:- match: { prefix: "/" }route: { cluster: local_cluster }- name: web_service_2domains: ["*.magedu.com",“magedu.com"]routes:- match: { prefix: "/" }redirect:host_redirect: "www.ik8s.io"http_filters:- name: envoy.filters.http.routertyped_config:"@type": type.googleapis.com/envoy.extensions.filters.http.router.v3.Routerclusters:- name: local_clusterconnect_timeout: 0.25stype: STATIClb_policy: ROUND_ROBINload_assignment:cluster_name: local_clusterendpoints:- lb_endpoints:- endpoint:address:socket_address: { address: 192.168.1.61, port_value: 8080 }- endpoint:address:socket_address: { address: 192.168.1.62, port_value: 8080 }
3.1.2 cat docker-compose.yaml
$ cat docker-compose.yaml
version: '3.3'services:envoy:image: envoyproxy/envoy:v1.23.4volumes:- ./envoy.yaml:/etc/envoy/envoy.yamlnetworks:envoymesh:ipv4_address: 192.168.1.60aliases:- front-proxydepends_on:- webserver01- webserver02webserver01:image: ikubernetes/demoapp:v1.0environment:- PORT=8080hostname: webserver01networks:envoymesh:ipv4_address: 192.168.1.61aliases:- webserver01webserver02:image: ikubernetes/demoapp:v1.0environment:- PORT=8080hostname: webserver02networks:envoymesh:ipv4_address: 192.168.1.62aliases:- webserver02networks:envoymesh:driver: bridgeipam:config:- subnet: 192.168.1.0/24
3.1.3 启动服务
$ docker-compose up -d
Creating network "http-front-proxy_envoymesh" with driver "bridge"
Creating http-front-proxy_webserver01_1 ... done
Creating http-front-proxy_webserver02_1 ... done
Creating http-front-proxy_envoy_1 ... done
3.1.4 检查进程
$ docker-compose psName Command State Ports
-----------------------------------------------------------------------------------
http-front-proxy_envoy_1 /docker-entrypoint.sh envo ... Up 10000/tcp
http-front-proxy_webserver01_1 /bin/sh -c python3 /usr/lo ... Up
http-front-proxy_webserver02_1 /bin/sh -c python3 /usr/lo ... Up
3.1.5 访问测试
$ curl -H "host: www.ik8s.io" 192.168.1.60
iKubernetes demoapp v1.0 !! ClientIP: 192.168.1.60, ServerName: webserver01, ServerIP: 192.168.1.61!
$ curl -H "host: www.ik8s.io" 192.168.1.60
iKubernetes demoapp v1.0 !! ClientIP: 192.168.1.60, ServerName: webserver01, ServerIP: 192.168.1.61!
$ curl -H "host: www.ik8s.io" 192.168.1.60
iKubernetes demoapp v1.0 !! ClientIP: 192.168.1.60, ServerName: webserver02, ServerIP: 192.168.1.62!
$ curl -H "host: www.ik8s.io" 192.168.1.60
iKubernetes demoapp v1.0 !! ClientIP: 192.168.1.60, ServerName: webserver02, ServerIP: 192.168.1.62!
$ curl -H "host: www.ik8s.io" 192.168.1.60
iKubernetes demoapp v1.0 !! ClientIP: 192.168.1.60, ServerName: webserver02, ServerIP: 192.168.1.62!
$ curl -H "host: www.ik8s.io" 192.168.1.60
iKubernetes demoapp v1.0 !! ClientIP: 192.168.1.60, ServerName: webserver02, ServerIP: 192.168.1.62!
$ curl -H "host: www.ik8s.io" 192.168.1.60
iKubernetes demoapp v1.0 !! ClientIP: 192.168.1.60, ServerName: webserver01, ServerIP: 192.168.1.61!
3.1.6 查看服务
$ curl -I -H "host: www.ik8s.io" 192.168.1.60
HTTP/1.1 200 OK
content-type: text/html; charset=utf-8
content-length: 101
server: envoy
date: Thu, 30 May 2024 07:44:09 GMT
x-envoy-upstream-service-time: 4
3.1.7 停止
$ docker-compose down
Stopping http-front-proxy_envoy_1 ... done
Stopping http-front-proxy_webserver02_1 ... done
Stopping http-front-proxy_webserver01_1 ... done
Removing http-front-proxy_envoy_1 ... done
Removing http-front-proxy_webserver02_1 ... done
Removing http-front-proxy_webserver01_1 ... done
Removing network http-front-proxy_envoymesh
4 http-ingress(入向流量代理)
ll http-ingress/
total 12
-rw-rw-r-- 1 root root 581 May 30 14:43 docker-compose.yaml
-rw-rw-r-- 1 root root 1214 May 30 14:43 envoy.yaml
-rw-rw-r-- 1 root root 425 May 30 14:43 README.md
4.1.1 cat envoy.yaml
# cat envoy.yaml
static_resources:listeners:- name: listener_0address:socket_address: { address: 0.0.0.0, port_value: 80 }filter_chains:- filters:- name: envoy.filters.network.http_connection_manager # 用于处理HTTP流量的HTTP连接管理器过滤器typed_config:"@type": type.googleapis.com/envoy.extensions.filters.network.http_connection_manager.v3.HttpConnectionManagerstat_prefix: ingress_http # 指定统计前缀codec_type: AUTO # 使用自动检测的编解码器类型route_config: # 路由配置name: local_route # 路由名称virtual_hosts: # 虚拟主机- name: web_service_1 # 虚拟主机名称domains: ["*"] # 匹配所有域名routes: # 路由规则- match: { prefix: "/" } # 匹配所有route: { cluster: local_cluster } # 凡是请求/的域名,都路由到local_clusterhttp_filters: # 定义HTTP过滤器- name: envoy.filters.http.router # 过滤器名称,这是一个基本的路由过滤器typed_config:"@type": type.googleapis.com/envoy.extensions.filters.http.router.v3.Routerclusters: # 集群配置- name: local_cluster # 集群名称connect_timeout: 0.25s # 连接超时时间type: STATIC # 集群类型:静态lb_policy: ROUND_ROBIN # 负载均衡策略:轮询load_assignment:cluster_name: local_cluster # 集群名称endpoints: # 集群端点- lb_endpoints: # 被轮询的端点- endpoint: address:socket_address: { address: 127.0.0.1, port_value: 8080 } # 由于业务pod中还有一个envoy proxy,而流量都是先打到envoy proxy上的,然后再由envoy proxy代理请求到业务pod,所以这里是127.0.0.1:8080(因为都在同一个pod中)
4.1.2 cat docker-compose.yaml
$ cat docker-compose.yaml
version: '3'services:envoy:image: envoyproxy/envoy:v1.23.4volumes:- ./envoy.yaml:/etc/envoy/envoy.yamlenvironment:- ENVOY_UID=0- ENVOY_GID=0networks:envoymesh:ipv4_address: 192.168.1.60aliases:- ingresswebserver01:image: ikubernetes/demoapp:v1.0environment:- PORT=8080- HOST=127.0.0.1network_mode: "service:envoy"depends_on:- envoynetworks:envoymesh:driver: bridgeipam:config:- subnet: 192.168.1.0/24
4.1.3 启动容器
$ docker-compose up -d
Creating network "http-ingress_envoymesh" with driver "bridge"
Creating http-ingress_envoy_1 ... done
Creating http-ingress_webserver01_1 ... done$ docker-compose ps Name Command State Ports
-------------------------------------------------------------------------------
http-ingress_envoy_1 /docker-entrypoint.sh envo ... Up 10000/tcp
http-ingress_webserver01_1 /bin/sh -c python3 /usr/lo ... Up
4.1.4 测试
$ curl 192.168.1.60
iKubernetes demoapp v1.0 !! ClientIP: 127.0.0.1, ServerName: d835488efd6a, ServerIP: 192.168.1.60!
4.1.5 进入容器中检查
# id
uid=0(root) gid=0(root) groups=0(root)
4.1.6 停止
$ docker-compose down
Stopping http-ingress_webserver01_1 ... done
Stopping http-ingress_envoy_1 ... done
Removing http-ingress_webserver01_1 ... done
Removing http-ingress_envoy_1 ... done
Removing network http-ingress_envoymesh
5.http-egress(出向流量代理)
ll http-egress/
total 12
-rw-rw-r-- 1 root root 961 May 30 14:43 docker-compose.yaml
-rw-rw-r-- 1 root root 1386 May 30 14:43 envoy.yaml
-rw-rw-r-- 1 root root 593 May 30 14:43 README.md
5.1.1 cat envoy.yaml
$ cat envoy.yaml
# Author: MageEdu <mage@magedu.com>
# Site: www.magedu.com
static_resources:listeners:- name: listener_0address:socket_address: { address: 127.0.0.1, port_value: 80 }filter_chains:- filters:- name: envoy.filters.network.http_connection_managertyped_config:"@type": type.googleapis.com/envoy.extensions.filters.network.http_connection_manager.v3.HttpConnectionManagerstat_prefix: ingress_httpcodec_type: AUTOroute_config:name: local_routevirtual_hosts:- name: web_service_1domains: ["*"]routes:- match: { prefix: "/" }route: { cluster: web_cluster }http_filters:- name: envoy.filters.http.routertyped_config:"@type": type.googleapis.com/envoy.extensions.filters.http.router.v3.Routerclusters:- name: web_clusterconnect_timeout: 0.25stype: STATIClb_policy: ROUND_ROBINload_assignment:cluster_name: web_clusterendpoints:- lb_endpoints:- endpoint:address:socket_address: { address: 192.168.1.61, port_value: 80 }- endpoint:address:socket_address: { address: 192.168.1.62, port_value: 80 }
5.1.2 cat docker-compose.yaml
$ cat docker-compose.yaml
# Author: MageEdu <mage@magedu.com>
# Site: www.magedu.com
version: '3.3'services:envoy:image: envoyproxy/envoy:v1.23.4volumes:- ./envoy.yaml:/etc/envoy/envoy.yamlnetworks:envoymesh:ipv4_address: 192.168.1.60aliases:- front-proxydepends_on:- webserver01- webserver02client:image: ikubernetes/admin-toolbox:v1.0network_mode: "service:envoy"depends_on:- envoywebserver01:image: ikubernetes/demoapp:v1.0hostname: webserver01networks:envoymesh:ipv4_address: 192.168.1.61aliases:- webserver01webserver02:image: ikubernetes/demoapp:v1.0hostname: webserver02networks:envoymesh:ipv4_address: 192.168.1.62aliases:- webserver02networks:envoymesh:driver: bridgeipam:config:- subnet: 192.168.1.0/24
5.1.3 docker pull
$ docker pull ikubernetes/admin-toolbox:v1.0
v1.0: Pulling from ikubernetes/admin-toolbox
c9b1b535fdd9: Already exists
1de2c4c6c672: Pull complete
Digest: sha256:620435ec75aee84cfaf40e75c7466d5a4d10e71f5884ccc5cf7a475538fdc937
Status: Downloaded newer image for ikubernetes/admin-toolbox:v1.0
docker.io/ikubernetes/admin-toolbox:v1.0
5.1.4 启动容器与检查
$ docker-compose up -d
Creating network "http-egress_envoymesh" with driver "bridge"
Creating http-egress_webserver02_1 ... done
Creating http-egress_webserver01_1 ... done
Creating http-egress_envoy_1 ... done
Creating http-egress_client_1 ... done$ docker-compose psName Command State Ports
------------------------------------------------------------------------------
http-egress_client_1 /bin/sh -c sleep 999999 Up
http-egress_envoy_1 /docker-entrypoint.sh envo ... Up 10000/tcp
http-egress_webserver01_1 /bin/sh -c python3 /usr/lo ... Up
http-egress_webserver02_1 /bin/sh -c python3 /usr/lo ... Up
5.1.5 访问测试
$ docker ps | grep egress
970b9d4ea168 ikubernetes/admin-toolbox:v1.0 "/bin/sh -c 'sleep 9…" About a minute ago Up About a minute http-egress_client_1
b1e4094c1c05 envoyproxy/envoy:v1.23.4 "/docker-entrypoint.…" About a minute ago Up About a minute 10000/tcp http-egress_envoy_1
92d446d8fc5c ikubernetes/demoapp:v1.0 "/bin/sh -c 'python3…" About a minute ago Up About a minute http-egress_webserver01_1
c01fb7e252c2 ikubernetes/demoapp:v1.0 "/bin/sh -c 'python3…" About a minute ago Up About a minute http-egress_webserver02_1$ docker exec -it http-egress_client_1 /bin/sh
[root@b1e4094c1c05 /]# curl 127.0.0.1
iKubernetes demoapp v1.0 !! ClientIP: 192.168.1.60, ServerName: webserver01, ServerIP: 192.168.1.61!
[root@b1e4094c1c05 /]# curl 127.0.0.1
iKubernetes demoapp v1.0 !! ClientIP: 192.168.1.60, ServerName: webserver01, ServerIP: 192.168.1.61!
[root@b1e4094c1c05 /]# curl 127.0.0.1
iKubernetes demoapp v1.0 !! ClientIP: 192.168.1.60, ServerName: webserver02, ServerIP: 192.168.1.62!
[root@b1e4094c1c05 /]# curl 127.0.0.1
5.1.6 清理容器
$ docker-compose down
Stopping http-egress_client_1 ... done
Stopping http-egress_envoy_1 ... done
Stopping http-egress_webserver01_1 ... done
Stopping http-egress_webserver02_1 ... done
Removing http-egress_client_1 ... done
Removing http-egress_envoy_1 ... done
Removing http-egress_webserver01_1 ... done
Removing http-egress_webserver02_1 ... done
Removing network http-egress_envoymesh
6 admin接口
Administration interface — envoy 1.23.12-9689bc documentationhttps://www.envoyproxy.io/docs/envoy/v1.23.12/operations/admin.html#administration-interface
Envoy内建了一个管理服务(administration server),它支持查询和修改操作。
但是它本身没有任何安全机制,有可能暴露私有数据(例如统计数据、集群名称和证书信息等),因此非常有必要精心编排其访问控制机制以避免非授权访问。
admin:access_log: [] # 访问日志协议的相关配置,通常需要指定日志过滤器及日志配置等;access_log_path: # 管理接口的访问日志文件路径,无须记录访问日志时可使用/dev/null;profile_path: # cpu profile 的输出路径,默认为/var/log/envoy/envoy.profaddress: # 监听的套接字;socket_address:protocol:address:port_value:# 配置示例
admin:access_log: /tmp/admin_access.logaddress:socket_address: { address: 0.0.0.0, port_value: 9901 } # 此处仅为了方便测试,才监听0.0.0.0,正常来说,应该使用127.0.0.1,或者人为的把9901加上认证机制。
ll admin-interface/
total 12
-rw-rw-r-- 1 foot foot 917 May 30 14:43 docker-compose.yaml
-rw-rw-r-- 1 foot foot 1832 May 30 14:43 envoy.yaml
-rw-rw-r-- 1 foot foot 713 May 30 14:43 README.md
6.1.1 cat envoy.yaml
$ cat envoy.yaml
# profile_path和access_log_path如果不需要的话,一定要配置为/dev/null
admin:profile_path: /tmp/envoy.profaccess_log_path: /tmp/admin_access.log
#
admin:profile_path: /tmp/envoy.profaccess_log_path: /tmp/admin_access.logaddress:socket_address:address: 0.0.0.0port_value: 9901static_resources:listeners:- name: listener_0address:socket_address: { address: 0.0.0.0, port_value: 80 }filter_chains:- filters:- name: envoy.filters.network.http_connection_managertyped_config:"@type": type.googleapis.com/envoy.extensions.filters.network.http_connection_manager.v3.HttpConnectionManagerstat_prefix: ingress_httpcodec_type: AUTOroute_config:name: local_routevirtual_hosts:- name: web_service_1domains: ["*.ik8s.io", "ik8s.io"]routes:- match: { prefix: "/" }route: { cluster: local_cluster }- name: web_service_2domains: ["*.magedu.com",“magedu.com"]routes:- match: { prefix: "/" }redirect:host_redirect: "www.ik8s.io"http_filters:- name: envoy.filters.http.routertyped_config:"@type": type.googleapis.com/envoy.extensions.filters.http.router.v3.Routerclusters:- name: local_clusterconnect_timeout: 0.25stype: STATIClb_policy: ROUND_ROBINload_assignment:cluster_name: local_clusterendpoints:- lb_endpoints:- endpoint:address:socket_address: { address: 192.168.1.61, port_value: 8080 }- endpoint:address:socket_address: { address: 192.168.1.62, port_value: 8080 }
6.1.2 cat docker-compose.yaml
$ cat docker-compose.yaml
admin:profile_path: /tmp/envoy.profaccess_log_path: /tmp/admin_access.logaddress:socket_address:address: 0.0.0.0port_value: 9901
version: '3.3'services:envoy:image: envoyproxy/envoy:v1.23.4volumes:- ./envoy.yaml:/etc/envoy/envoy.yamlenvironment:- ENVOY_UID=0- ENVOY_GID=0networks:envoymesh:ipv4_address: 192.168.1.60aliases:- front-proxydepends_on:- webserver01- webserver02webserver01:image: ikubernetes/demoapp:v1.0environment:- PORT=8080hostname: webserver01networks:envoymesh:ipv4_address: 192.168.1.61aliases:- webserver01webserver02:image: ikubernetes/demoapp:v1.0environment:- PORT=8080hostname: webserver02networks:envoymesh:ipv4_address: 192.168.1.62aliases:- webserver02networks:envoymesh:driver: bridgeipam:config:- subnet: 192.168.1.0/24
6.1.3 启动与检查
$ docker-compose up -d
Creating network "admin-interface_envoymesh" with driver "bridge"
Creating admin-interface_webserver01_1 ... done
Creating admin-interface_webserver02_1 ... done
Creating admin-interface_envoy_1 ... done$ docker-compose psName Command State Ports
----------------------------------------------------------------------------------
admin-interface_envoy_1 /docker-entrypoint.sh envo ... Up 10000/tcp
admin-interface_webserver01_1 /bin/sh -c python3 /usr/lo ... Up
admin-interface_webserver02_1 /bin/sh -c python3 /usr/lo ... Up
6.1.4 访问测试
$ curl 192.168.1.60:9901/help
admin commands are:/: Admin home page/certs: print certs on machine/clusters: upstream cluster status/config_dump: dump current Envoy configs (experimental)/contention: dump current Envoy mutex contention stats (if enabled)/cpuprofiler: enable/disable the CPU profiler/drain_listeners: drain listeners/healthcheck/fail: cause the server to fail health checks/healthcheck/ok: cause the server to pass health checks/heapprofiler: enable/disable the heap profiler/help: print out list of admin commands/hot_restart_version: print the hot restart compatibility version/listeners: print listener info/logging: query/change logging levels/memory: print current allocation/heap usage/quitquitquit: exit the server/ready: print server state, return 200 if LIVE, otherwise return 503/reopen_logs: reopen access logs/reset_counters: reset all counters to zero/runtime: print runtime values/runtime_modify: modify runtime values/server_info: print server version/status information/stats: print server stats/stats/prometheus: print server stats in prometheus format/stats/recentlookups: Show recent stat-name lookups/stats/recentlookups/clear: clear list of stat-name lookups and counter/stats/recentlookups/disable: disable recording of reset stat-name lookup names/stats/recentlookups/enable: enable recording of reset stat-name lookup names
6.1.5 常用的接口
/config_dump # GET 打印Envoy加载的各类配置信息,支持include_eds、master和resource等查询参数
/listners # GET 列出所有侦听器,支持使用“GET /listners?format=json”
/clusters # GET 额外支持使用 "GET /clusters?format=json"
/stats # 按需输出统计数据,例如 GET /stats?filter=regex,另外还支持json和prometheus格式
/stats/prometheus # 输出Prometheus统计信息
/drain_listners # POST 驱逐所有listner,支持使用inboundonly(仅入站侦听器)和graceful(优雅关闭)等查询参数;
/runtime # GET 以json格式输出所有运行时相关值
/runtime_modify # POST 修改运行时参数接口之一
/server_info
6.1.6 停止
$ docker-compose down
Stopping admin-interface_envoy_1 ... done
Stopping admin-interface_webserver02_1 ... done
Stopping admin-interface_webserver01_1 ... done
Removing admin-interface_envoy_1 ... done
Removing admin-interface_webserver02_1 ... done
Removing admin-interface_webserver01_1 ... done
Removing network admin-interface_envoymesh
7 运行时接口
Runtime — envoy 1.31.0-dev-e543e1 documentationhttps://www.envoyproxy.io/docs/envoy/latest/configuration/operations/runtime.html
7.1.1 知识背景
(1)相较于静态资源配置来说,xDS API的动态配置机制使得Envoy的配置系统极具弹性;但有时候配置的变动仅需要修改个别的功能特性,若通过xDS接口完成未免有些动静过大,Runtime便是面向这种场景的配置接口;
Runtime就是一个虚拟文件系统树,可通过一至多个本地文件系统目录、静态资源、RTDS动态发现和Admin Interface进行定文和配置;
每个配置称为一个Layer,因而也称为"Layered Runtime",这些Layer最终叠加生效;(2)换句话说,Runtime是与Envoy一起部署的外置实时配置系统,用于支持更改配置设置而无需重启Envov或更改主配置;运行时配置相关的运行时参数也称为“功能标志 (feature flags)” 或“决策者 (decider)” ;
通过运行时参数更改配置将实时生效;(3)运行时配置的实现也称为运行时配置供应者;Envoy当前支持的运行时配置的实现是由多个层级组成的虚拟文件系统;
Envoy在配置的目录中监视符号链接的交换空间,并在发生交换时重新加载文件树;
但Envoy会使用默认运行时值和“null”提供给程序以确保其正确运行,因此,运行时配置系统并不必不可少;
7.1.2 配置运行时
启用Envov的运行时配置机制,需要在Bootstrap文件中予以启用和配置;
- 定义在bootstrap配置文件中的layered_runtime顶级字段之下;
- 一旦在bootstrap中给出layered_runtime字段,则至少要定义出一个layer;# 示例
layered_runtime: # 运行配置供应者,未指定时则使用null供应者,即所有参数均加载其默认值;layers: # 运行时的层级列表,后面的层将覆盖先前层上的配置;- name: static_layer_0 # 运行时的层级名称,仅用于“GET/runtime”时的输出;static_layer: {...} # 静态运行时层级,遵循运行时probobuf JSON编码格式,不同于静态的xDS资源,静态运行时层一样可被后面的层所覆盖;disk_layer: {...} # 基于本地磁盘的运行时层级;symlink_root: ... # 通过符号链接访问的文件系统树;subdirectory: ... # 指定要在根目录中加载的子目录;append_service_cluster: # 是否将服务集群附加至符号链接根目录下的子路径上;admin_layer: {...} # 管理控制台运行时层级,即通过/runtime管理端点查看,通过/runtime_modify管理端点修改的配置方式;rtds_layer: {...} # 运行时发现服务(runtime discovery service)层级,即通过xDS API中的RTDS API动态发现相关的层级配置;name: ... # 在rtds_config上为RTDS层订阅的资源;rtds_config: ... # RTDS的ConfigSource
$ ll layered-runtime/
total 12
-rw-rw-r-- 1 root root 861 May 30 14:43 docker-compose.yaml
-rw-rw-r-- 1 root root 2002 May 30 14:43 envoy.yaml
-rw-rw-r-- 1 root root 576 May 30 14:43 README.md
7.1.3 cat envoy.yaml
$ cat envoy.yaml admin:profile_path: /tmp/envoy.profaccess_log_path: /tmp/admin_access.logaddress:socket_address:address: 0.0.0.0port_value: 9901layered_runtime:layers:- name: static_layer_0static_layer:health_check:min_interval: 5- name: admin_layer_0admin_layer: {}static_resources:listeners:- name: listener_0address:socket_address: { address: 0.0.0.0, port_value: 80 }filter_chains:- filters:- name: envoy.filters.network.http_connection_managertyped_config:"@type": type.googleapis.com/envoy.extensions.filters.network.http_connection_manager.v3.HttpConnectionManagerstat_prefix: ingress_httpcodec_type: AUTOroute_config:name: local_routevirtual_hosts:- name: web_service_1domains: ["*.ik8s.io", "ik8s.io"]routes:- match: { prefix: "/" }route: { cluster: local_cluster }- name: web_service_2domains: ["*.magedu.com",“magedu.com"]routes:- match: { prefix: "/" }redirect:host_redirect: "www.ik8s.io"http_filters:- name: envoy.filters.http.routertyped_config:"@type": type.googleapis.com/envoy.extensions.filters.http.router.v3.Routerclusters:- name: local_clusterconnect_timeout: 0.25stype: STATIClb_policy: ROUND_ROBINload_assignment:cluster_name: local_clusterendpoints:- lb_endpoints:- endpoint:address:socket_address: { address: 192.168.1.61, port_value: 8080 }- endpoint:address:socket_address: { address: 192.168.1.62, port_value: 8080 }
7.1.4 cat docker-compose.yaml
$ cat docker-compose.yaml
version: '3.3'services:envoy:image: envoyproxy/envoy:v1.23.4volumes:- ./envoy.yaml:/etc/envoy/envoy.yamlnetworks:envoymesh:ipv4_address: 192.168.1.60aliases:- front-proxydepends_on:- webserver01- webserver02webserver01:image: ikubernetes/demoapp:v1.0environment:- PORT=8080hostname: webserver01networks:envoymesh:ipv4_address: 192.168.1.61aliases:- webserver01webserver02:image: ikubernetes/demoapp:v1.0environment:- PORT=8080hostname: webserver02networks:envoymesh:ipv4_address: 192.168.1.62aliases:- webserver02networks:envoymesh:driver: bridgeipam:config:- subnet: 192.168.1.0/24
7.1.5 启动与检查
$ docker-compose up -d
Creating network "layered-runtime_envoymesh" with driver "bridge"
Creating layered-runtime_webserver01_1 ... done
Creating layered-runtime_webserver02_1 ... done
Creating layered-runtime_envoy_1 ... done$ docker-compose psName Command State Ports
----------------------------------------------------------------------------------
layered-runtime_envoy_1 /docker-entrypoint.sh envo ... Up 10000/tcp
layered-runtime_webserver01_1 /bin/sh -c python3 /usr/lo ... Up
layered-runtime_webserver02_1 /bin/sh -c python3 /usr/lo ... Up
7.1.6 查看默认数据
$ curl 192.168.1.60:9901/runtime
{"layers": ["static_layer_0","admin_layer_0"],"entries": {"health_check.min_interval": {"final_value": "5","layer_values": ["5",""]}}
}
7.1.7 修改超时时间
$ curl -XPOST 192.168.1.60:9901/runtime_modify?health_check.min_interval=15
OK$ curl 192.168.1.60:9901/runtime
{"layers": ["static_layer_0","admin_layer_0"],"entries": {"health_check.min_interval": {"final_value": "15","layer_values": ["5","15"]}}
}
参考文档连接
原文链接:https://blog.csdn.net/qq_42515722/article/details/132363590