(六)服务调用-Ribbon

经过以上的学习,已经实现了服务的注册和服务发现。当启动某个服务的时候,可以通过HTTP的形式将信息注册到注册中心,并且可以通过SpringCloud提供的工具获取注册中心的服务列表。但是服务之间的调用还存在很多的问题,如何更加方便的调用微服务,多个微服务的提供者如何选择,如何负载均衡等

1.Ribbon概述

1.1什么是Ribbon

是 Netflixfa 发布的一个负载均衡器,有助于控制 HTTP 和 TCP客户端行为。在 SpringCloud 中,Eureka一般配合Ribbon进行使用,Ribbon提供了客户端负载均衡的功能,Ribbon利用从Eureka中读取到的服务信息,在调用服务节点提供的服务时,会合理的进行负载。

在SpringCloud中可以将注册中心和Ribbon配合使用,Ribbon自动的从注册中心中获取服务提供者的列表信息,并基于内置的负载均衡算法,请求服务。

架构说明:
(六)服务调用-Ribbon
Ribbon在工作时分成两步
第一步先选择 EurekaServer ,它优先选择在同一个区域内负载较少的server.
第二步再根据用户指定的策略,在从server取到的服务注册列表中选择一个地址。
其中Ribbon提供了多种策略:比如轮询、随机和根据响应时间加权

1.2Ribbon的主要作用

(1)服务调用
基于Ribbon实现服务调用, 是通过拉取到的所有服务列表组成(服务名-请求路径的)映射关系。借助RestTemplate 最终进行调用。

(2)负载均衡
当有多个服务提供者时,Ribbon可以根据负载均衡的算法自动的选择需要调用的服务地址

2.案例应用

之前写样例时候没有引入spring-cloud-starter-ribbon也可以使用ribbon

<dependency>
    <groupId>org.springframework.cloud</groupId>
    <artifactId>spring-cloud-starter-netflix-ribbon</artifactId>
</dependency>

猜测spring-cloud-starter-netflix-eureka-client自带了spring-cloud-starter-ribbon引用
证明如下: 可以看到spring-cloud-starter-netflix-eureka-client 确实引入了Ribbon
(六)服务调用-Ribbon

服务提供者

cloud-provider-payment-8001
cloud-provider-payment-8002

pom

        <!--eureka-client-->
        <dependency>
            <groupId>org.springframework.cloud</groupId>
            <artifactId>spring-cloud-starter-netflix-eureka-client</artifactId>
        </dependency>

        <dependency>
            <groupId>org.springframework.boot</groupId>
            <artifactId>spring-boot-starter-web</artifactId>
        </dependency>
        <dependency>
            <groupId>org.springframework.boot</groupId>
            <artifactId>spring-boot-starter-actuator</artifactId>
        </dependency>
        <dependency>
            <groupId>org.mybatis.spring.boot</groupId>
            <artifactId>mybatis-spring-boot-starter</artifactId>
        </dependency>

主启动类

@SpringBootApplication
@EnableEurekaClient
@EnableDiscoveryClient
@MapperScan(basePackages = "com.mine.springcloud.payment.dao")
public class PaymentApplication8001 {

    public static void main(String[] args) {
        SpringApplication.run(PaymentApplication8001.class, args);
    }
}

Controller

@RestController
@Slf4j
public class PaymentController {

    @Resource
    private PaymentService paymentService;

    @Resource
    private DiscoveryClient discoveryClient;

    @Value(value = "${server.port}")
    private Integer serverPort;

    @PostMapping(value = "/payment/create")
    public CommonResult create(@RequestBody Payment payment) {
        int result = paymentService.create(payment);
        log.info("*****插入操作返回结果:" + result);
        if (result > 0) {
            return CommonResult.init(200, "create success");
        } else {
            return CommonResult.init(500, "create fials");
        }

    }

    @GetMapping(value = "payment/get/{id}")
    public CommonResult<Payment> getPaymentById(@PathVariable("id") Long id) {
        Payment payment = paymentService.getPaymentById(id);
        log.info("*****查询结果:{}", payment);
        if (payment != null) {
            return CommonResult.init(200, "get success serverPort:" + serverPort, payment);
        } else {
            return CommonResult.init(404, "get null");
        }
    }
}

服务消费者

cloud-consumer-order-80

pom

        <!--eureka-client-->
        <dependency>
            <groupId>org.springframework.cloud</groupId>
            <artifactId>spring-cloud-starter-netflix-eureka-client</artifactId>
        </dependency>

        <dependency>
            <groupId>org.springframework.boot</groupId>
            <artifactId>spring-boot-starter-web</artifactId>
        </dependency>
        <dependency>
            <groupId>org.springframework.boot</groupId>
            <artifactId>spring-boot-starter-actuator</artifactId>
        </dependency>

主启动类

@SpringBootApplication
@EnableEurekaClient
@EnableDiscoveryClient
public class OrderApplication {

    public static void main(String[] args) {
        SpringApplication.run(OrderApplication.class, args);
    }
}

Controller

@RestController
@Slf4j
public class OrderController {

    //private String PAYMENT_URL = "http://localhost:8001/";

    private String PAYMENT_SERVICE_URL = "http://CLOUD-PAYMENT-SERVICE";

    @Resource
    private RestTemplate restTemplate;

    @Resource
    private DiscoveryClient discoveryClient;

    @PostMapping(value = "/consumer/payment/create")
    public CommonResult create(Payment payment) {
        return restTemplate.postForObject(PAYMENT_SERVICE_URL + "/payment/create", payment, CommonResult.class);
    }

    @GetMapping(value = "/consumer/payment/get/{id}")
    public CommonResult<Payment> getPaymentById(@PathVariable("id") Long id) {
        return restTemplate.getForObject(PAYMENT_SERVICE_URL + "/payment/get/" + id, CommonResult.class, id);
    }
}

配置Bean

@Configuration
public class ApplicationContextConfig {

    @Bean
    @LoadBalanced
    public RestTemplate restTemplate() {
        return new RestTemplate();
    }
}

3.负载均衡概述

3.1什么是负载均衡

在搭建网站时,如果单节点的 web服务性能和可靠性都无法达到要求;或者是在使用外网服务时,经常担心被人攻破,一不小心就会有打开外网端口的情况,通常这个时候加入负载均衡就能有效解决服务问题
负载均衡是一种基础的网络服务,其原理是通过运行在前面的负载均衡服务,按照指定的负载均衡算法,将流量分配到后端服务集群上,从而为系统提供并行扩展的能力。
负载均衡的应用场景包括流量包、转发规则以及后端服务,由于该服务有内外网个例、健康检查等功能,能够有效提供系统的安全性和可用性。
(六)服务调用-Ribbon

3.2客户端负载均衡与服务端负载均衡

服务端负载均衡
先发送请求到负载均衡服务器或者软件,然后通过负载均衡算法,在多个服务器之间选择一个进行访问;即在服务器端再进行负载均衡算法分配

客户端负载均衡
客户端会有一个服务器地址列表,在发送请求前通过负载均衡算法选择一个服务器,然后进行访问,这是客户端负载均衡;即在客户端就进行负载均衡算法分配

4.Ribbon负载均衡策略

Ribbon内置了多种负载均衡策略,内部负责复杂均衡的*接口为com.netflix.loadbalancer.IRule ,实现方式如下
(六)服务调用-Ribbon

  • com.netflix.loadbalancer.RoundRobinRule :以轮询的方式进行负载均衡。
  • com.netflix.loadbalancer.RandomRule :随机策略
  • com.netflix.loadbalancer.RetryRule :重试策略。
  • com.netflix.loadbalancer.WeightedResponseTimeRule :权重策略。会计算每个服务的权重,越高的被调用的可能性越大。
  • com.netflix.loadbalancer.BestAvailableRule :最佳策略。遍历所有的服务实例,过滤掉故障实例,并返回请求数最小的实例返回。
  • com.netflix.loadbalancer.AvailabilityFilteringRule :可用过滤策略。过滤掉故障和请求数超过阈值的服务实例,再从剩下的实力中轮询调用。

4.1在服务消费者的application.yml配置文件中修改负载均衡策略

##需要调用的微服务名称
cloud-order-service:
 ribbon:
  NFLoadBalancerRuleClassName: com.netflix.loadbalancer.RandomRule

4.2在服务消费者配置类中修改负载均衡策略

新建包com.mine.springcloud.myrule与主启动类所扫描包区分开
定义负载均衡策略

import com.netflix.loadbalancer.IRule;
import com.netflix.loadbalancer.RandomRule;
import org.springframework.context.annotation.Bean;
import org.springframework.context.annotation.Configuration;

@Configuration
public class MySelfRule {

    @Bean
    public IRule myRule() {
        return new RandomRule();//定义为随机
    }
}

主启动类添加@RibbonClient注解

@SpringBootApplication
@EnableEurekaClient
@EnableDiscoveryClient
@RibbonClient(name = "CLOUD-PAYMENT-SERVICE", configuration = MySelfRule.class)
public class OrderApplication {

    public static void main(String[] args) {
        SpringApplication.run(OrderApplication.class, args);
    }
}

4.3自定义本地负载均衡算法

  • ApplicationContextBean去掉注解@LoadBalanced
@Configuration
public class ApplicationContextBean
{
    @Bean
    //@LoadBalanced
    public RestTemplate getRestTemplate()
    {
        return new RestTemplate();
    }
}
  • 主启动类去掉注释@RibbonClient
@SpringBootApplication
@EnableEurekaClient
@EnableDiscoveryClient
//@RibbonClient(name = "CLOUD-PAYMENT-SERVICE", configuration = MySelfRule.class)
public class OrderApplication {

    public static void main(String[] args) {
        SpringApplication.run(OrderApplication.class, args);
    }
}
  • LoadBalancer接口
package com.mine.springcloud.order.lb;

import org.springframework.cloud.client.ServiceInstance;

import java.util.List;

public interface LoadBalancer {

    ServiceInstance instances(List<ServiceInstance> list);
}
  • MyLB
package com.mine.springcloud.order.lb;

import lombok.extern.slf4j.Slf4j;
import org.springframework.cloud.client.ServiceInstance;
import org.springframework.stereotype.Component;

import java.util.List;
import java.util.concurrent.atomic.AtomicInteger;

@Component
@Slf4j
public class MyLB implements LoadBalancer {

    private AtomicInteger atomicInteger = new AtomicInteger(0);

    private Integer getAndIncrement() {
        int current;
        int next;
        do {
            current = atomicInteger.get();
            next = current >= Integer.MAX_VALUE ? 0 : current + 1;
            log.info("next:" + next);
        } while (!atomicInteger.compareAndSet(current, next));
        return next;
    }

    @Override
    public ServiceInstance instances(List<ServiceInstance> list) {
        return list.get(getAndIncrement() % list.size());
    }
}
  • Controller
package com.mine.springcloud.order.controller;

import com.mine.springcloud.commons.base.CommonResult;
import com.mine.springcloud.commons.entity.Payment;
import com.mine.springcloud.order.lb.LoadBalancer;
import lombok.extern.slf4j.Slf4j;
import org.springframework.cloud.client.ServiceInstance;
import org.springframework.cloud.client.discovery.DiscoveryClient;
import org.springframework.http.ResponseEntity;
import org.springframework.web.bind.annotation.GetMapping;
import org.springframework.web.bind.annotation.PathVariable;
import org.springframework.web.bind.annotation.PostMapping;
import org.springframework.web.bind.annotation.RestController;
import org.springframework.web.client.RestTemplate;

import javax.annotation.Resource;
import java.util.List;

@RestController
@Slf4j
public class OrderController {

    //private String PAYMENT_URL = "http://localhost:8001/";

    private String PAYMENT_SERVICE_URL = "http://CLOUD-PAYMENT-SERVICE";

    @Resource
    private RestTemplate restTemplate;

    @Resource
    private LoadBalancer loadBalancer;

    @Resource
    private DiscoveryClient discoveryClient;

    @GetMapping(value = "/consumer/payment/lb")
    public CommonResult<Payment> getLB() {
        List<ServiceInstance> instances = discoveryClient.getInstances("CLOUD-PAYMENT-SERVICE");
        if (instances == null || instances.size() == 0) {
            return CommonResult.init(500, "instances is null");
        }
        ServiceInstance instance = loadBalancer.instances(instances);
        return restTemplate.getForObject(instance.getUri() + "/payment/lb", CommonResult.class);
    }
}

4.4原理

负载均衡算法:rest接口第几次请求数 % 服务器集群总数量 = 实际调用服务器位置下标 ,每次服务重启动后rest接口计数从1开始。
List instances = discoveryClient.getInstances("CLOUD-PAYMENT-SERVICE");
如:
List [0] instances = 127.0.0.1:8002
List [1] instances = 127.0.0.1:8001
8001+ 8002 组合成为集群,它们共计2台机器,集群总数为2, 按照轮询算法原理:
当总请求数为1时: 1 % 2 =1 对应下标位置为1 ,则获得服务地址为127.0.0.1:8001
当总请求数位2时: 2 % 2 =0 对应下标位置为0 ,则获得服务地址为127.0.0.1:8002
当总请求数位3时: 3 % 2 =1 对应下标位置为1 ,则获得服务地址为127.0.0.1:8001
当总请求数位4时: 4 % 2 =0 对应下标位置为0 ,则获得服务地址为127.0.0.1:8002
如此类推......

5.Ribbon中负载均衡的源码解析

5.1Ribbon中的关键组件

(六)服务调用-Ribbon

  • ServerList:可以响应客户端的特定服务的服务器列表。
  • ServerListFilter:可以动态获得的具有所需特征的候选服务器列表的过滤器。
  • ServerListUpdater:用于执行动态服务器列表更新。
  • Rule:负载均衡策略,用于确定从服务器列表返回哪个服务器。
  • Ping:客户端用于快速检查服务器当时是否处于活动状态。
  • LoadBalancer:负载均衡器,负责负载均衡调度的管理。

5.2@LoadBalanced注解

使用Ribbon完成客户端负载均衡往往是从一个注解开始的

/**
* 基于Ribbon的服务调用与负载均衡
*/
@LoadBalanced
@Bean
public RestTemplate getRestTemplate() {
  return new RestTemplate();
}

这个注解的主要作用是什么呢,查看源码

/**
* Annotation to mark a RestTemplate bean to be configured to use a
LoadBalancerClient
* @author Spencer Gibb
*/
@Target({ ElementType.FIELD, ElementType.PARAMETER, ElementType.METHOD })
@Retention(RetentionPolicy.RUNTIME)
@Documented
@Inherited
@Qualifier
public @interface LoadBalanced {
}

通过注释可以知道@LoadBalanced注解是用来给RestTemplate做标记,方便我们对RestTemplate添加一个LoadBalancerClient,以实现客户端负载均衡。

5.3自动装配

根据SpringBoot中的自动装配规则可以在 spring-cloud-netflix-ribbon-2.1.0.RELEASE.jar 中可以找到 spring.factories

org.springframework.boot.autoconfigure.EnableAutoConfiguration=\org.springframew
ork.cloud.netflix.ribbon.RibbonAutoConfiguration

找到自动装配的类RibbonAutoConfiguration

@Configuration
@Conditional({RibbonAutoConfiguration.RibbonClassesConditions.class})
@RibbonClients
@AutoConfigureAfter(name = {"org.springframework.cloud.netflix.eureka.EurekaClientAutoConfiguration"})
@AutoConfigureBefore({LoadBalancerAutoConfiguration.class,
AsyncLoadBalancerAutoConfiguration.class})
@EnableConfigurationProperties({RibbonEagerLoadProperties.class,
ServerIntrospectorProperties.class})
public class RibbonAutoConfiguration {
  @Bean
  public SpringClientFactory springClientFactory() {
    SpringClientFactory factory = new SpringClientFactory();
    factory.setConfigurations(this.configurations);
    return factory;
  }
 
  @Bean
  @ConditionalOnMissingBean({LoadBalancerClient.class})
  public LoadBalancerClient loadBalancerClient() {
    return new RibbonLoadBalancerClient(this.springClientFactory());
  }
  //省略
}

通过 RibbonAutoConfiguration 引入了 LoadBalancerAutoConfiguration 配置类

5.4负载均衡调用

@Configuration
@ConditionalOnClass({RestTemplate.class})
@ConditionalOnBean({LoadBalancerClient.class})
@EnableConfigurationProperties({LoadBalancerRetryProperties.class})
public class LoadBalancerAutoConfiguration {
  @LoadBalanced
  @Autowired(required = false)
  private List<RestTemplate> restTemplates = Collections.emptyList();
  @Autowired( required = false)
  private List<LoadBalancerRequestTransformer> transformers = Collections.emptyList();

  public LoadBalancerAutoConfiguration() {
  }

  @Bean
  public SmartInitializingSingleton
  loadBalancedRestTemplateInitializerDeprecated(ObjectProvider<List<RestTemplateCustomizer>> restTemplateCustomizers) {
    return () -> {
      restTemplateCustomizers.ifAvailable((customizers) -> {
        Iterator var2 = this.restTemplates.iterator();
        while(var2.hasNext()) {
          RestTemplate restTemplate = (RestTemplate)var2.next();
          Iterator var4 = customizers.iterator();
          while(var4.hasNext()) {
            RestTemplateCustomizer customizer =(RestTemplateCustomizer)var4.next();
            customizer.customize(restTemplate);
          }
        }
     });
    };
  }

  @Bean
  @ConditionalOnMissingBean
  public LoadBalancerRequestFactory loadBalancerRequestFactory(LoadBalancerClient loadBalancerClient) {
    return new LoadBalancerRequestFactory(loadBalancerClient,this.transformers);
  }

  @Configuration
  @ConditionalOnClass({RetryTemplate.class})
  public static class RetryInterceptorAutoConfiguration {
    public RetryInterceptorAutoConfiguration() {
    }

    @Bean
    @ConditionalOnMissingBean
    public RetryLoadBalancerInterceptor ribbonInterceptor(LoadBalancerClient loadBalancerClient, LoadBalancerRetryProperties properties,LoadBalancerRequestFactory requestFactory, LoadBalancedRetryFactory loadBalancedRetryFactory) {
      return new RetryLoadBalancerInterceptor(loadBalancerClient,properties, requestFactory, loadBalancedRetryFactory);
    }

    @Bean
    @ConditionalOnMissingBean
    public RestTemplateCustomizer restTemplateCustomizer(RetryLoadBalancerInterceptor loadBalancerInterceptor) {
      return (restTemplate) -> {
        List<ClientHttpRequestInterceptor> list = new ArrayList(restTemplate.getInterceptors());
        list.add(loadBalancerInterceptor);
        restTemplate.setInterceptors(list);
     };
   }

  }

  @Configuration
  @ConditionalOnClass({RetryTemplate.class})
  public static class RetryAutoConfiguration {
    public RetryAutoConfiguration() {
    }
    @Bean
    @ConditionalOnMissingBean
    public LoadBalancedRetryFactory loadBalancedRetryFactory() {
      return new LoadBalancedRetryFactory() {
      };
    }
  }

  @Configuration
  @ConditionalOnMissingClass({"org.springframework.retry.support.RetryTemplate"})
  static class LoadBalancerInterceptorConfig {
    LoadBalancerInterceptorConfig() {
    }
    @Bean
    public LoadBalancerInterceptor ribbonInterceptor(LoadBalancerClient loadBalancerClient, LoadBalancerRequestFactory requestFactory) {
      return new LoadBalancerInterceptor(loadBalancerClient,requestFactory);
    }

    @Bean
    @ConditionalOnMissingBean
    public RestTemplateCustomizer restTemplateCustomizer(LoadBalancerInterceptor loadBalancerInterceptor) {
      return (restTemplate) -> {
        List<ClientHttpRequestInterceptor> list = new ArrayList(restTemplate.getInterceptors());
        list.add(loadBalancerInterceptor);
        restTemplate.setInterceptors(list);
      };
    }
  }
}

在该自动化配置类中,主要做了下面三件事:

  • 创建了一个 LoadBalancerInterceptor 的Bean,用于实现对客户端发起请求时进行拦截,以实现客户端负载均衡。
  • 创建了一个 RestTemplateCustomizer 的Bean,用于给 RestTemplate 增加LoadBalancerInterceptor 拦截器。
  • 维护了一个被 @LoadBalanced 注解修饰的 RestTemplate 对象列表,并在这里进行初始化,通过调用 RestTemplateCustomizer 的实例来给需要客户端负载均衡的 RestTemplate 增加

LoadBalancerInterceptor 拦截器。

public class LoadBalancerInterceptor implements ClientHttpRequestInterceptor {
  private LoadBalancerClient loadBalancer;
  private LoadBalancerRequestFactory requestFactory;
  public LoadBalancerInterceptor(LoadBalancerClient loadBalancer, LoadBalancerRequestFactory requestFactory) {
    this.loadBalancer = loadBalancer;
    this.requestFactory = requestFactory;
  }
  public LoadBalancerInterceptor(LoadBalancerClient loadBalancer) {
    this(loadBalancer, new LoadBalancerRequestFactory(loadBalancer));
  }
  public ClientHttpResponse intercept(HttpRequest request, byte[] body, ClientHttpRequestExecution execution) throws IOException {
    URI originalUri = request.getURI();
    String serviceName = originalUri.getHost();
    Assert.state(serviceName != null, "Request URI does not contain a valid hostname: " + originalUri);
    return (ClientHttpResponse)this.loadBalancer.execute(serviceName, this.requestFactory.createRequest(request, body, execution));
  }
}

通过源码以及之前的自动化配置类,我们可以看到在拦截器中注入了 LoadBalancerClient 的实现。当一个被 @LoadBalanced 注解修饰的 RestTemplate 对象向外发起HTTP请求时,会被LoadBalancerInterceptor 类的 intercept 函数所拦截。由于我们在使用RestTemplate时候采用了服务名作为host,所以直接从 HttpRequest 的URI对象中通过getHost()就可以拿到服务名,然后调用execute 函数去根据服务名来选择实例并发起实际的请求。

分析到这里, LoadBalancerClient 还只是一个抽象的负载均衡器接口,所以我们还需要找到它的具体实现类来进一步分析。通过查看ribbon的源码,我们可以很容易的在org.springframework.cloud.netflix.ribbon 包下找到对应的实现类:RibbonLoadBalancerClient 。

public class RibbonLoadBalancerClient implements LoadBalancerClient {
  public ServiceInstance choose(String serviceId) {
    Server server = this.getServer(serviceId);
    return server == null ? null : new RibbonLoadBalancerClient.RibbonServer(serviceId, server, this.isSecure(server, serviceId), this.serverIntrospector(serviceId).getMetadata(server));
  }
 
  public <T> T execute(String serviceId, LoadBalancerRequest<T> request) throws IOException {
    ILoadBalancer loadBalancer = this.getLoadBalancer(serviceId);
    Server server = this.getServer(loadBalancer);
    if (server == null) {
      throw new IllegalStateException("No instances available for " + serviceId);
    } else {
      RibbonLoadBalancerClient.RibbonServer ribbonServer = new RibbonLoadBalancerClient.RibbonServer(serviceId, server, this.isSecure(server, serviceId), this.serverIntrospector(serviceId).getMetadata(server));
      return this.execute(serviceId, ribbonServer, request);
    }
  }
 
  protected Server getServer(String serviceId) {
    return this.getServer(this.getLoadBalancer(serviceId));
  }
 
  protected Server getServer(ILoadBalancer loadBalancer) {
    return loadBalancer == null ? null : loadBalancer.chooseServer("default");
  }
 
  protected ILoadBalancer getLoadBalancer(String serviceId) {
    return this.clientFactory.getLoadBalancer(serviceId);
  }
 
  //省略...
}
  • ServiceInstance choose(String serviceId):根据传入的服务id,从负载均衡器中为指定的服务选择一个服务实例。
  • T execute(String serviceId, LoadBalancerRequest request):根据传入的服务id,指定的负载均衡器中的服务实例执行请求。
  • T execute(String serviceId, ServiceInstance serviceInstance, LoadBalancerRequest request):根据传入的服务实例,执行请求。

从 RibbonLoadBalancerClient 代码可以看出,实际负载均衡的是通过 ILoadBalancer 来实现的

@Bean
@ConditionalOnMissingBean
public ILoadBalancer ribbonLoadBalancer(IClientConfig config, ServerList<Server>serverList, ServerListFilter<Server> serverListFilter, IRule rule, IPing ping,ServerListUpdater serverListUpdater) {
  return (ILoadBalancer)(this.propertiesFactory.isSet(ILoadBalancer.class, this.name) ?(ILoadBalancer)this.propertiesFactory.get(ILoadBalancer.class, config, this.name) : new ZoneAwareLoadBalancer(config, rule, ping, serverList, serverListFilter,serverListUpdater));
}
上一篇:Spring Cloud Sleuth 分布式请求链路跟踪


下一篇:4、负载均衡:Ribbon、Feign