想对看过的spring cloud 各个项目进行一个简单的搭建,并使用docker部署,目前包含的项目有Eureka、Ribbon、Feign、Hystrix、Hystrix Dashboard、Turbine聚合监控、Zuul、SpringBootAdmin等Spring Config 、OAuth2未集成进来,但是在我的主页里有单独的实例,后续会慢慢都集成进来。
github地址:https://github.com/liangxiaobo/springbootcloud-all 分支 master-docker
项目模块:
├── client-common-dependencys
├── client-feign
├── client-gateway-zuul
├── client-order-ribbon
├── client-turbine-monitor
├── docker-compose-base.yml
├── docker-compose.yml
├── docker-start.sh
├── docker-zipkin-shell.sh
├── eureka-server
├── mvn-package-docker.sh
├── pom.xml
├── service-common-dependencys
├── service-order
├── service-user
├── spring-boot-admin-server
├── springbootdemoall.iml
├── src
├── swagger-doc
└── var
项目 | 端口 | 描述 |
---|---|---|
eureka-server | 8761 | 服务的注册与发现 |
service-user | 8763 | 服务提供者 |
service-order | 8764 | 服务提供者 |
client-feign | 8765 | 负载均衡 Feign |
client-order-ribbon | 8766 | 负载均衡 Ribbon |
client-gateway-zuul | 8771 | 网关 Zuul |
client-turbine-monitor | 8767 | Turbine聚合监控 |
swagger-doc | 8772 | 生成在线Api |
spring-boot-admin-server | 8773 | spring-boot-admin-server |
项目中看到两个项目 service-common-dependencys和client-common-dependencys是pom项目,用来测试统一管理引用依赖的项目(分别是service项目和client项目的父类);
当父类pom用dependencies来管理依赖的话,那么子项目必须继承所有依赖
当父类pom用dependencyManagement来管理依赖的话,那么子项目不是必须继承,而是有选择的引用依赖;我这里使用的是前者
eureka-server 项目 参照 spring cloud 搭建集群Eureka Server
spring-boot-admin-server 参照 SpringBoot Admin 2.0
client-roder-ribbon 参照 使用RestTemplate和Ribbon来消费服务和Hystrix熔断功能
service-user、service-order 就是简单的web项目
Feign项目 client-feign
Feign是一个声明式的Web Service客户端,它使得编写Web Serivce客户端变得更加简单。我们只需要使用Feign来创建一个接口并用注解来配置它既可完成。它具备可插拔的注解支持,包括Feign注解和JAX-RS注解。Feign也支持可插拔的编码器和解码器。Spring Cloud为Feign增加了对Spring MVC注解的支持,还整合了Ribbon和Eureka来提供均衡负载的HTTP客户端实现。
依赖pom.xml
org.springframework.cloud
spring-cloud-starter-openfeign
pom中引用了父类 ,继承父类的所有依赖,其中有Hystrix、admin-client、eureka等
com.spring.nahong.client.common.dependency
client-common-dependencys
${all.version}
../client-common-dependencys
看一下启动类
@SpringBootApplication
@EnableEurekaClient
@EnableFeignClients
@EnableHystrixDashboard
public class ClientFeignApplication {
public static void main(String[] args) {
SpringApplication.run(ClientFeignApplication.class, args);
}
/**
* springboot 版本如果是2.0则需要添加 ServletRegistrationBean
* 因为springboot的默认路径不是 "/hystrix.stream",
* 只要在自己的项目里配置上下面的servlet就可以了
* 第一次访问hystrix.stream 会出现 Unable to connect to Command Metric Stream
*/
@Bean
public ServletRegistrationBean getServlet() {
HystrixMetricsStreamServlet streamServlet = new HystrixMetricsStreamServlet();
ServletRegistrationBean registrationBean = new ServletRegistrationBean(streamServlet);
registrationBean.setLoadOnStartup(1);
registrationBean.addUrlMappings("/hystrix.stream");
registrationBean.setName("HystrixMetricsStreamServlet");
return registrationBean;
}
}
@EnableFeignClients 开启了feign @EnableHystrixDashboard开启了熔断hystrix的界面
定义Feign接口UserClientFeign
@FeignClient(value = "service-user", configuration = FeignConfig.class, fallback = UserHystrix.class)
public interface UserClientFeign {
@GetMapping("/user/say")
String sayFromClient(@RequestParam("name") String name);
}
上面声明了指向service-user服务 UserHystrix是对应的熔断类
@Component
public class UserHystrix implements UserClientFeign {
@Override
public String sayFromClient(String name) {
return "hi , sorry error!";
}
}
定义一个服务UserService
@Service
public class UserService {
@Autowired
UserClientFeign userClientFeign;
public String say(String name) {
return userClientFeign.sayFromClient(name);
}
}
声明一个Feign的配置类FeignConfig
@Configuration
public class FeignConfig {
@Bean
public Retryer feignRetryer() {
return new Retryer.Default(100, SECONDS.toMillis(1), 5);
}
}
创建一个UserController
@RestController
public class UserController {
@Autowired
UserService userService;
@RequestMapping("/user/hi")
public String say(@RequestParam("name") String name) {
return userService.say(name);
}
}
启用Feign的hystrix最重要的是在配置文件中添加
feign:
hystrix:
enabled: true
可以启动 eureka-server service-user client-feign 访问http://localhost:8765/user/hi?name=liajngbo
如果测试负载均衡功能可以多运行一个service-user用不同的端口8863,然后在浏览器刷新 http://localhost:8765/user/hi?name=liajngbo 会发现有不同的结果
Hi, my name is liajngbo, port: 8763
Hi, my name is liajngbo, port: 8863
client-feign项目中已经配置了hystrix,现在访问http://localhost:8765/hystrix
在输入框中输入http://localhost:8765/hystrix.stream点击按扭
Turbine项目 client-turbine-monitor
Turbine能够汇集监控信息,并将聚合后的信息提供给Hystrix Dashboard来集中展示和监控。
项目配置起来相当简单,只是2.0版本会有点小坑
pom.xml全部代码请参阅源码(源码包含了admin-client的依赖)这只是turbine的依赖:
org.springframework.boot
spring-boot-starter-actuator
org.springframework.cloud
spring-cloud-starter-netflix-turbine
org.springframework.cloud
spring-cloud-starter-netflix-hystrix-dashboard
启动类:
@SpringBootApplication
@EnableTurbine
@EnableHystrixDashboard
public class ClientTurbineMonitorApplication {
public static void main(String[] args) {
SpringApplication.run(ClientTurbineMonitorApplication.class, args);
}
}
@EnableTurbine 开启了turbine功能,@EnableHystrixDashboard开启了hystrix监控UI
一般的配置文件:
turbine:
app-config: client-feign,client-order-ribbon,client-gateway-zuul # 指定了要监控的应用名字
cluster-name-expression: new String("default") # 表示集群的名字为default
combine-host-port: true # 表示同一主机上的服务通过host和port的组合来进行区分,默认情况下是使用host来区分,这样会使本地调试有问题
这里会有一个问题,turbine会去访问其它项目的/hystrix.stream路径,默认情况下turbine访问的是/actuator/hystrix.stream,这个路径访问不通,因为其它项目注册的路由地址是/hystrix.stream,所以要纠正turbine项目访问的默认地址,在配置文件中添加turbine.instanceUrlSuffix=hystrix.stream,management.context-path=/
配置文件:
management:
context-path: /
turbine:
app-config: client-feign,client-order-ribbon,client-gateway-zuul # 指定了要监控的应用名字
cluster-name-expression: new String("default") # 表示集群的名字为default
combine-host-port: true # 表示同一主机上的服务通过host和port的组合来进行区分,默认情况下是使用host来区分,这样会使本地调试有问题
instanceUrlSuffix: hystrix.stream
测试一下turbine,必须启动多个项目 eureka-server、service-user、service-order、client-feign、client-order-ribbon、client-gateway-zuul、client-turbine-monitor
访问turbine之前要先访问一下其它项目,不然不会有数据显示turbine会处于loading中
访问 http://localhost:8767/turbine.stream 会有json格式的数据流
{"rollingCountFallbackSuccess":0,"rollingCountFallbackFailure":0,"propertyValue_circuitBreakerRequestVolumeThreshold":20,"propertyValue_circuitBreakerForceOpen":false,"propertyValue_metricsRollingStatisticalWindowInMilliseconds":10000,"latencyTotal_mean":10,"rollingMaxConcurrentExecutionCount":0,"type":"HystrixCommand","rollingCountResponsesFromCache":0,"rollingCountBadRequests":0,"rollingCountTimeout":0,"propertyValue_executionIsolationStrategy":"THREAD","rollingCountFailure":0,"rollingCountExceptionsThrown":0,"rollingCountFallbackMissing":0,"threadPool":"service-user","latencyExecute_mean":10,"isCircuitBreakerOpen":false,"errorCount":0,"rollingCountSemaphoreRejected":0,"group":"service-user","latencyTotal":{"0":10,"99":10,"100":10,"25":10,"90":10,"50":10,"95":10,"99.5":10,"75":10},"requestCount":0,"rollingCountCollapsedRequests":0,"rollingCountShortCircuited":0,"propertyValue_circuitBreakerSleepWindowInMilliseconds":5000,"latencyExecute":{"0":10,"99":10,"100":10,"25":10,"90":10,"50":10,"95":10,"99.5":10,"75":10},"rollingCountEmit":0,"currentConcurrentExecutionCount":0,"propertyValue_executionIsolationSemaphoreMaxConcurrentRequests":10,"errorPercentage":0,"rollingCountThreadPoolRejected":0,"propertyValue_circuitBreakerEnabled":true,"propertyValue_executionIsolationThreadInterruptOnTimeout":true,"propertyValue_requestCacheEnabled":true,"rollingCountFallbackRejection":0,"propertyValue_requestLogEnabled":true,"rollingCountFallbackEmit":0,"rollingCountSuccess":0,"propertyValue_fallbackIsolationSemaphoreMaxConcurrentRequests":10,"propertyValue_circuitBreakerErrorThresholdPercentage":50,"propertyValue_circuitBreakerForceClosed":false,"name":"UserClientFeign#sayFromClient(String)","reportingHosts":1,"propertyValue_executionIsolationThreadPoolKeyOverride":"null","propertyValue_executionIsolationThreadTimeoutInMilliseconds":1000,"propertyValue_executionTimeoutInMilliseconds":1000}
data: {"currentCorePoolSize":10,"currentLargestPoolSize":2,"propertyValue_metricsRollingStatisticalWindowInMilliseconds":10000,"currentActiveCount":0,"currentMaximumPoolSize":10,"currentQueueSize":0,"type":"HystrixThreadPool","currentTaskCount":2,"currentCompletedTaskCount":2,"rollingMaxActiveThreads":0,"rollingCountCommandRejections":0,"name":"service-user","reportingHosts":1,"currentPoolSize":2,"propertyValue_queueSizeRejectionThreshold":5,"rollingCountThreadsExecuted":0}
访问hystrix UI http://localhost:8767/hystrix
网关Zuul client-gateway-zuul
项目依赖,完整的pom.xml中包含了其它的依赖
org.springframework.cloud
spring-cloud-starter-netflix-zuul
启动类:
@SpringBootApplication
@EnableEurekaClient
@EnableZuulProxy
@EnableHystrixDashboard
public class ClientGatewayZuulApplication {
public static void main(String[] args) {
SpringApplication.run(ClientGatewayZuulApplication.class, args);
}
/**
* springboot 版本如果是2.0则需要添加 ServletRegistrationBean
* 因为springboot的默认路径不是 "/hystrix.stream",
* 只要在自己的项目里配置上下面的servlet就可以了
* 第一次访问hystrix.stream 会出现 Unable to connect to Command Metric Stream
*
*/
@Bean
public ServletRegistrationBean getServlet() {
HystrixMetricsStreamServlet streamServlet = new HystrixMetricsStreamServlet();
ServletRegistrationBean registrationBean = new ServletRegistrationBean(streamServlet);
registrationBean.setLoadOnStartup(1);
registrationBean.addUrlMappings("/hystrix.stream");
registrationBean.setName("HystrixMetricsStreamServlet");
return registrationBean;
}
}
开启网关zuul @EnableZuulProxy
@EnableEurekaClient 开启eureka客户端
@EnableHystrixDashboard Hystrix Dashboard
主要配置:
zuul:
routes:
hiapi:
path: /hiapi/**
serviceId: service-user
ribbonapi:
path: /ribbonapi/**
serviceId: client-order-ribbon
feignapi:
path: /feignapi/**
serviceId: client-feign
prefix: /v1
routes下面是路由,hiapi、ribbonapi、feignapi是自定义路由名称,下面path是路由规则,serviceId指定服务名称,prefix给路由加上前缀
注意:因为项目是我写好的demo,其中包含了hystrix功能,在/hystrix中Thread Pools会一直处于loading状态,这是由于Zuul默认会使用信号量来实现隔离,只有通过Hystrix配置把隔离机制改成为线程池的方式才能够得以展示,
SEMAPHORE - 它在调用线程上执行,并发请求受信号量计数的限制(Zuul默认此策略)
THREAD - 它在一个单独的线程上执行,并发请求受到线程池中线程数的限制
配置文件中添加
thread-pool:
use-separate-thread-pools: true
ribbon-isolation-strategy: thread # 每个路由使用独立的线程池
Zuul中提供Filter的作用有哪些,我觉得分为如下几点:
- 网关是暴露在外面的,必须要进行权限控制
- 可以针对服务做控制,在路由的时候处理,比如服务降级
- 防止爬虫,利用Filter对请求进行过滤
- 流量控制,只允许最高的并发量,保护后端的服务
- 灰度发布,可以针对不用的用户进行路由来实现灰度
Filter种类
- pre:可以在请求被路由之前调用
- route:在路由请求时候被调用
- post:在route和error过滤器之后被调用
- error:处理请求时发生错误时被调用
zuul的Filter的生命周期,见下图:
创建一个Filter类MyFilter,filterType类型为"pre":
@Component
public class MyFilter extends ZuulFilter {
private static Logger logger = LoggerFactory.getLogger(MyFilter.class);
@Override
public String filterType() {
return FilterConstants.PRE_TYPE; // pre
}
@Override
public int filterOrder() {
return 0;
}
@Override
public boolean shouldFilter() {
return true;
}
@Override
public Object run() throws ZuulException {
RequestContext ctx = RequestContext.getCurrentContext();
HttpServletRequest request = ctx.getRequest();
Object accessToken = request.getParameter("token");
if (accessToken == null) {
logger.warn("token is empty");
ctx.setSendZuulResponse(false);
ctx.setResponseStatusCode(401);
try {
ctx.getResponse().getWriter().write("token is empty");
}catch (Exception e) {
return null;
}
}
logger.info("ok");
return null;
}
}
shouldFilter是决定这个过滤器需不需要执行,返回false则不执行,这个也可以利用配置中心来做,达到动态的开启关闭效果,
filterOrder是表示过滤器执行的顺序,数字越小,优先级越高
run里面就是我们自己要执行的业务逻辑,这里我们验证了一下接口参数中有没有token
zuul自身是集成了hystrix的,所以它带有熔断功能
创建一个ServiceUserFallbackProvider类实现接口FallbackProvider,实现service-user的熔断功能:
@Component
public class ServiceUserFallbackProvider implements FallbackProvider {
static final Logger logger = LoggerFactory.getLogger(ServiceUserFallbackProvider.class);
@Override
public String getRoute() {
// 表明是为哪个微服务提供回退,*表示为所有微服务提供回退,当前只为service-user提供
return "service-user";
}
@Override
public ClientHttpResponse fallbackResponse(String route, Throwable cause) {
logger.info("route: " + route);
return new ClientHttpResponse() {
@Override
public HttpStatus getStatusCode() throws IOException {
return HttpStatus.OK;
}
@Override
public int getRawStatusCode() throws IOException {
return 200;
}
@Override
public String getStatusText() throws IOException {
return "OK";
}
@Override
public void close() {
}
@Override
public InputStream getBody() throws IOException {
return new ByteArrayInputStream("oooops! error,i'm the fallback.".getBytes());
}
@Override
public HttpHeaders getHeaders() {
HttpHeaders headers = new HttpHeaders();
headers.setContentType(MediaType.APPLICATION_JSON);
return headers;
}
};
}
}
启动eureka-server、service-user、service-order、client-feign、client-order-ribbon、client-gateway-zuul,访问http://localhost:8771/v1/hiapi/user/say?name=liang&token=123
正常结果显示:
Hi, my name is liang, port: 8763
去掉token访问:http://localhost:8771/v1/hiapi/user/say?name=liang
token is empty
测试熔断效果,关停service-user项目 访问: http://localhost:8771/v1/hiapi/user/say?name=liang&token=123
oooops! error,i'm the fallback.
使用docker部署项目
Docker是一个用于开发、交付和运行应用的开放平台,Docker被设计用于更快地交付应用。Docker可以将应用程序和基础设施层隔离,并且可以将基础设施当作程序一样进行管理。使用Docker,可以更快地打包代码、测试以及部署,并且可以减少从编写到部署运行代码的周期。
Docker将内核容器特性(LXC)、工作流和工具集成,以帮助管理和部署应用。
什么是Docker
核心是,Docker是一种在安全隔离的容器中运行近乎所有应用的方式,这种隔离性和安全性允许你在同一个主机上同时运行多个容器,而容器的这种轻量级特性,无需消耗运行hpervisor所需的额外负载,意味着你可以节省更多的硬件资源。
docker如何安装这里不讲,自行解决
这里演示在单台宿主机上的测试结果
因为项目是我提前写好的项目,可以看到项目中的pom文件中有:
com.spotify
docker-maven-plugin
${docker.plugin.version}
${docker.image.prefix}/${project.artifactId}
src/main/docker
/
${project.build.directory}
${project.build.finalName}.jar
latest
${docker.image.prefix}是我在父类中定义的liangwang
imageName是镜像名称
dockerDirectory是Dockerfile路径
resources是资源的配置,jar所在的目录
imageTags是镜像的tag版本的意思,image:[tag]
Dockerfile
FROM java
VOLUME /tmp
ADD eureka-server-0.0.1-SNAPSHOT.jar app.jar
#RUN bash -c 'touch /app.jar'
ENTRYPOINT ["java","-Djava.security.egd=file:/dev/./urandom","-jar","-Dspring.profiles.active=${SPRING_PROFILES_ACTIVE}","/app.jar"]
EXPOSE 8761 # 开放的端口
EXPOSE 是开放的端口 多个端口可以以空格隔开 "8761 8762 8763"
ENTRYPOINT 是容器启动后执行的命令
springboot打包的项目jar的运行命令 "java -jar app.jar" 如果启用不同的配置文件的话要 “java -jar app.jar --spring.profiles.active=test”,所以上面定义了一个变量来接收容器启动设置的参数SPRING_PROFILES_ACTIVE 容器启动时加 -e "SPRING_PROFILES_ACTIVE=test"
maven 打包 docker 命令
mvn clean package docker:build -Dmaven.test.skip=true
# -Dmaven.test.skip 跳过单元测试
容器启动:
docker run -d -e "SPRING_PROFILES_ACTIVE=test-peer1" --name eureka-server -p 8761:8761 -it liangwang/eureka-server
同一台宿主机可以使用容器名称来通信 比如 eureka-server容器使用link参数将其它链接到容器2
docker run -d -e "SPRING_PROFILES_ACTIVE=test" --name service-user-01 --link=eureka-server -p 8763:8763 -it liangwang/service-user
对于注册与发现服务,eureka-server和其它eureka-client中的 serviceUri.defaultZone=http://eureka-server:8761/eureka/
例举各别:
eureka-server的配置文件
spring:
profiles: test-peer1
application:
name: eureka-server
server:
port: 8761
eureka:
instance:
prefer-ip-address: true
hostname: eureka-server
instance-id: ${spring.cloud.client.ipAddress}:${server.port}
client:
service-url:
defaultZone: http://eureka-server:8761/eureka/ # 注意客户端中也使用eureka-server
register-with-eureka: false
fetch-registry: false
下面可以按顺序打包项目:
项目根目录中有一个mvn-package-docker.sh
#!/bin/bash
echo "开始执行打包....[start]"
mvn clean
cd eureka-server
mvn package docker:build -Dmaven.test.skip=true
cd ../service-user
mvn package docker:build -Dmaven.test.skip=true
cd ../service-order
mvn package docker:build -Dmaven.test.skip=true
cd ../spring-boot-admin-server
mvn package docker:build -Dmaven.test.skip=true
cd ../client-feign
mvn package docker:build -Dmaven.test.skip=true
cd ../client-order-ribbon
mvn package docker:build -Dmaven.test.skip=true
cd ../client-gateway-zuul
mvn package docker:build -Dmaven.test.skip=true
cd ../client-turbine-monitor
mvn package docker:build -Dmaven.test.skip=true
cd ../swagger-doc
mvn package docker:build -Dmaven.test.skip=true
echo "执行打包结束.....[end]"
也可以手动打包各别你想打包的项目
打包完之后,执行命令 docker images
[root@swarm03 ~]# docker images
REPOSITORY TAG IMAGE ID CREATED SIZE
liangwang/swagger-doc latest 9d477baa8695 2 days ago 689MB
liangwang/client-turbine-monitor latest bb26593b2683 2 days ago 692MB
liangwang/client-gateway-zuul latest a2148967a806 2 days ago 693MB
liangwang/client-order-ribbon latest 71a24105d743 2 days ago 693MB
liangwang/client-feign latest 6ae21d5dab6d 2 days ago 693MB
liangwang/spring-boot-admin-server latest 6d9314544f13 2 days ago 697MB
liangwang/service-order latest 48b8fd2e0de4 2 days ago 691MB
liangwang/service-user latest f33c40c5b79a 2 days ago 691MB
liangwang/eureka-server latest f5e714d47a6f 2 days ago 691MB
java latest d23bdf5b1b1b 20 months ago 643MB
除java之外都是打包生成的镜像
下面要运行容器来跑这些镜像,项目根目录有一个docker-start.sh脚本,里面是跑各个容器的命令,最好不要执行这个脚本,因为没有时间停顿,所以容器全部运行没有前后顺序,如果eureka-server没有跑起来,可能其它项目会报错,最好先手动一条一条执行,这里是演示测试所以没有用容器编排
# eureka-server docker 启动命令:
docker run -d -e "SPRING_PROFILES_ACTIVE=test-peer1" --name eureka-server -p 8761:8761 -it liangwang/eureka-server
# service-user-01 docker 启动命令:
docker run -d -e "SPRING_PROFILES_ACTIVE=test" --name service-user-01 --link=eureka-server -p 8763:8763 -it liangwang/service-user
# service-user-02 docker 启动命令:
docker run -d -e "SPRING_PROFILES_ACTIVE=test" --name service-user-02 --link=eureka-server -p 8863:8863 -it liangwang/service-user --server.port=8863
# service-order-01 docker 启动命令:
docker run -d -e "SPRING_PROFILES_ACTIVE=test" --name service-order-01 --link=eureka-server -p 8764:8764 -it liangwang/service-order
# service-order-02 docker 启动命令:
docker run -d -e "SPRING_PROFILES_ACTIVE=test" --name service-order-02 --link=eureka-server -p 8864:8864 -it liangwang/service-order --server.port=8864
# spring-boot-admin-server docker 启动命令:
docker run -d -e "SPRING_PROFILES_ACTIVE=test" --name spring-boot-admin-server --link=eureka-server -p 8773:8773 -it liangwang/spring-boot-admin-server
# client-feign docker 启动命令:
docker run -d -e "SPRING_PROFILES_ACTIVE=test" --name client-feign --link=eureka-server -p 8765:8765 -it liangwang/client-feign
# client-order-ribbon docker 启动命令:
docker run -d -e "SPRING_PROFILES_ACTIVE=test" --name client-order-ribbon --link=eureka-server -p 8766:8766 -it liangwang/client-order-ribbon
# client-gateway-zuul docker 启动命令:
docker run -d -e "SPRING_PROFILES_ACTIVE=test" --name client-gateway-zuul --link=eureka-server -p 8771:8771 -it liangwang/client-gateway-zuul
# client-turbine-monitor docker 启动命令:
docker run -d -e "SPRING_PROFILES_ACTIVE=test" --name client-turbine-monitor --link=eureka-server -p 8767:8767 -it liangwang/client-turbine-monitor
# swagger-doc docker 启动命令:
docker run -d -e "SPRING_PROFILES_ACTIVE=test" --name swagger-doc --link=eureka-server -p 8772:8772 -it liangwang/swagger-doc
没有问题的情况下,执行docker ps能看到正在运行的容器进程
[root@swarm03 ~]# docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
0c70c83fd5f2 liangwang/client-turbine-monitor "java -Djava.secur..." 2 days ago Up 2 days 0.0.0.0:8767->8767/tcp client-turbine-monitor
1b78c139bded liangwang/client-gateway-zuul "java -Djava.secur..." 2 days ago Up 2 days 0.0.0.0:8771->8771/tcp client-gateway-zuul
1582be4ac882 liangwang/client-order-ribbon "java -Djava.secur..." 2 days ago Up 2 days 0.0.0.0:8766->8766/tcp client-order-ribbon
d13d2fdf24ae liangwang/client-feign "java -Djava.secur..." 2 days ago Up 2 days 0.0.0.0:8765->8765/tcp client-feign
3f149a4873cf liangwang/spring-boot-admin-server "java -Djava.secur..." 2 days ago Up 2 days 0.0.0.0:8773->8773/tcp spring-boot-admin-server
399150b6fa6e liangwang/service-order "java -Djava.secur..." 2 days ago Up 2 days 8764/tcp, 0.0.0.0:8864->8864/tcp service-order-02
e7ffefe94d3e liangwang/service-order "java -Djava.secur..." 2 days ago Up 2 days 0.0.0.0:8764->8764/tcp service-order-01
9be1cabe9b5a liangwang/service-user "java -Djava.secur..." 2 days ago Up 2 days 8763/tcp, 0.0.0.0:8863->8863/tcp service-user-02
7b82a04ddbeb liangwang/service-user "java -Djava.secur..." 2 days ago Up 2 days 0.0.0.0:8763->8763/tcp service-user-01
ee0fdbee9f33 liangwang/eureka-server "java -Djava.secur..." 2 days ago Up 2 days 0.0.0.0:8761->8761/tcp eureka-server
这时候访问 http://localhost:8761,我项目部署在测试服务器所以我的访问地址为 http://172.16.10.177:8761/
能看到项目都已经注册进来了
看admin-server 访问http://localhost:8773,我的访问地址为 http://172.16.10.177:8773
改造项目增加zipkin链路追踪
使用镜像运行zipkin项目
请先拉取镜像
docker pull openzipkin/zipkin
# 运行容器
# docker run -d -p 9411:9411 -e MYSQL_USER=root -e MYSQL_PASS=password -e MYSQL_HOST=192.168.0.8 -e STORAGE_TYPE=mysql openzipkin/zipkin
docker run -d -p 9411:9411 --name zipkin openzipkin/zipkin
在需要追踪的项目中添加依赖
org.springframework.cloud
spring-cloud-starter-zipkin
在application.yml中添加
spring:
zipkin:
base-url: http://localhost:9411 # 如果zipkin服务在其它服务器,localhost应为对应的IP
sleuth:
sampler:
probability: 1.0
sleuth.sampler.probability 是监控的百分比,默认的是0.1表示10%,这里给1.0表示全部监控
spring.zipkin.base-url:是zipkin-server的服务路径
访问 http://localhost:9411
后续我会集成OAuth2.0 JWT和Config配置中心
Docker Swarm 搭建SpringCloud集群