SpringCloud之超时重试

亦凉 2023-06-29 07:25 18阅读 0赞

[附：SpringCloud之系列教程汇总跳转地址][SpringCloud]

### Zuul的超时重试 ###

application.properties文件

zuul.retryable=true
    
    ribbon.ConnectTimeout=250
    ribbon.ReadTimeout=1000
    ribbon.OkToRetryOnAllOperations=false
    ribbon.MaxAutoRetries=1
    ribbon.MaxAutoRetriesNextServer=2

ZuulServiceFallbackProvider

package com.yj.zuul.provider;
    
    import java.io.ByteArrayInputStream;
    import java.io.IOException;
    import java.io.InputStream;
    import org.slf4j.Logger;
    import org.slf4j.LoggerFactory;
    import org.springframework.cloud.netflix.zuul.filters.route.ZuulFallbackProvider;
    import org.springframework.http.HttpHeaders;
    import org.springframework.http.HttpStatus;
    import org.springframework.http.MediaType;
    import org.springframework.http.client.ClientHttpResponse;
    import org.springframework.stereotype.Component;
    
    @Component
    public class ZuulServiceFallbackProvider implements ZuulFallbackProvider {
    	protected final Logger logger = LoggerFactory.getLogger(ZuulServiceFallbackProvider.class);
    
    	@Override
    	public String getRoute() {
    		// 服务id,可以用* 或者 null 代表所有服务都过滤
    		return "*";
    	}
    
    	@Override
    	public ClientHttpResponse fallbackResponse() {
    		return new ClientHttpResponse() {
    			@Override
    			public HttpHeaders getHeaders() {
    				HttpHeaders headers = new HttpHeaders();
    				headers.setContentType(MediaType.APPLICATION_JSON_UTF8);
    				return headers;
    			}
    
    			@Override
    			public InputStream getBody() throws IOException {
    				logger.info("服务不可用,请稍后再试...");
    				return new ByteArrayInputStream("服务不可用,请稍后再试...".getBytes());
    			}
    
    			@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() {
    
    			}
    		};
    	}
    }

验证，我们在Ribbon项目的接口中进行模拟超时的动作

System.out.println("i am hiRestTemplate,port:" + port+","+DateFormatUtils.format(new Date(), "yyyyMMdd HH:mm:ss"));
    TimeUnit.SECONDS.sleep(5);

然后从Zuul网关发起对Ribbon项目接口的GET方法的调用，发现Ribbon集群的两台实例，分别输出

i am hiRestTemplate2,port:8005,20200111 12:36:21
    i am hiRestTemplate2,port:8005,20200111 12:36:22
    i am hiRestTemplate2,port:8005,20200111 12:36:25
    i am hiRestTemplate2,port:8005,20200111 12:36:26

i am hiRestTemplate2,port:8015,20200111 12:36:23
    i am hiRestTemplate2,port:8015,20200111 12:36:24

超时重试的规则是这样的：

①首先负载到05端口的实例，发现超时，再调用一次（ribbon.MaxAutoRetries=1），就有了05实例21秒和22秒的两次调用日志

②发现还是没能响应，切换实例进行重试（ribbon.MaxAutoRetriesNextServer=2）,就有了15实例23秒和24秒的两次调用日志

③发现还是没能响应，切换实例进行重试（ribbon.MaxAutoRetriesNextServer=2）,就有了05实例25秒和26秒的两次调用日志

④上面的示例说的是Get请求，非Get请求则不会进行超时重试，因为我们设置了ribbon.OkToRetryOnAllOperations=false

重试的总次数为（MaxAutoRetries+1）\*（MaxAutoRetriesNextServer+1）= 6，如果6次后还是失败，触发ZuulServiceFallbackProvider中的熔断机制，输出“服务不可用,请稍后再试...”

### 微服务之间相互调用时的超时重试 ###

Ribbon加RestTemplate的方式

application.properties

hystrix.command.default.execution.timeout.enabled=true
    hystrix.command.default.execution.isolation.thread.timeoutInMilliseconds=5000
    
    spring.cloud.loadbalancer.retry.enabled=true
    
    ribbon.restclient.enabled=true
    ribbon.ConnectTimeout=250
    ribbon.ReadTimeout=1000
    ribbon.OkToRetryOnAllOperations=false
    ribbon.MaxAutoRetriesNextServer=2
    ribbon.MaxAutoRetries=1

RibbonService熔断方法

package com.yj.hystrix.ribbon.service;
    
    import org.springframework.beans.factory.annotation.Autowired;
    import org.springframework.stereotype.Service;
    import org.springframework.web.client.RestTemplate;
    import com.netflix.hystrix.contrib.javanica.annotation.HystrixCommand;
    
    @Service
    public class RibbonService {
    
    	@Autowired
    	private RestTemplate restTemplate;
    
    	@HystrixCommand(fallbackMethod = "ribbonError")
    	public String hiRibbon(String name) {
    		return restTemplate.getForObject("http://eureka-client/hiEureka?name=" + name, String.class);
    	}
    
    	public String ribbonError(String name) {
    		//降级,可以添加入库操作，后续可以根据这些记录手工恢复数据
    		System.out.println("restTemplate降级");
    		return "sorry " + name;
    	}
    }

Feign的方式

application.properties

feign.hystrix.enabled=true
    
    hystrix.command.default.execution.timeout.enabled=true
    hystrix.command.default.execution.isolation.thread.timeoutInMilliseconds=5000
    
    spring.cloud.loadbalancer.retry.enabled=true
    
    ribbon.ConnectTimeout=250
    ribbon.ReadTimeout=1000
    ribbon.OkToRetryOnAllOperations=false
    ribbon.MaxAutoRetriesNextServer=2
    ribbon.MaxAutoRetries=1

FeignServiceHystric熔断方法

package com.yj.hystrix.feign.service;
    
    import org.springframework.stereotype.Component;
    
    @Component
    public class FeignServiceHystric implements FeignService {
    	@Override
    	public String hiFeign(String name) {
    		// 降级,可以添加入库操作，后续可以根据这些记录手工恢复数据
    		System.out.println("feign降级");
    		return "sorry " + name;
    	}
    }

验证，我们在Eureka-Client项目的接口中进行模拟超时的动作

System.out.println("i am eureka client,port:" + port+","+DateFormatUtils.format(new Date(), "yyyy-MM-dd HH:mm:ss"));

然后从Hystrix发起对Eureka-Client项目接口的调用，发现Eureka-Client集群的两台实例，分别输出

i am eureka client,port:8004,2020-01-11 14:45:47
    i am eureka client,port:8004,2020-01-11 14:45:48
    i am eureka client,port:8004,2020-01-11 14:45:51
    i am eureka client,port:8004,2020-01-11 14:45:52

i am eureka client,port:8003,2020-01-11 14:45:49
    i am eureka client,port:8003,2020-01-11 14:45:50

超时重试的规则同上

[附：SpringCloud之系列教程汇总跳转地址][SpringCloud]

[SpringCloud]: https://blog.csdn.net/qq_24313635/article/details/104061217