Spring cloud integrates the Hystrix circuit breaker (shutdown and maintenance)

Posted by RockyShark on Thu, 20 Jan 2022 19:17:50 +0100

Note: stopping for further maintenance
Note: stopping for further maintenance
Note: stopping for further maintenance
Note: stopping for further maintenance
Note: stopping for further maintenance
Note: stopping for further maintenance
Say important things six times

Distributed problems:
Applications in complex distributed architecture have multiple dependencies, and each dependency will inevitably fail at some time, resulting in service avalanche

Official website information:

https://github.com/Netflix/Hystrix/wiki/How-To-Use

There are three main concepts of Hystrix:

Service degradation:
Conditions leading to degradation
1> Abnormal program operation: such as null pointer error, etc
2> Timeout
3> Service fuse triggers service degradation
4> Full thread pool / semaphore will also cause service degradation

Service fuse
When the maximum amount of access is reached, it directly denies access, and then calls the service degradation method to give friendly hints.
After the maximum number of accesses is reached, the service is degraded - > fusing - > recovering the calling link

Service current limiting
Second kill and other high concurrency operations are strictly prohibited. Overcrowding is prohibited. There are N in a second. Queue up and enter in an orderly manner

Service degradation control code construction

pom.xml import

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

Controller code

@HystrixCommand(fallbackMethod = "gets_TimeOutHandler",commandProperties = {
        @HystrixProperty(name="execution.isolation.thread.timeoutInMilliseconds",value = "3000")
})
@RequestMapping(value = "gets",method = RequestMethod.GET)
public String gets(){
    String result=null;
    try {
        TimeUnit.SECONDS.sleep(5);
        //todo business logic processing
        result= payService.getAll();
    } catch (InterruptedException e) {
        e.printStackTrace();
    } catch (Exception e) {
        return e.toString();
    }
    return result;
}

public String gets_TimeOutHandler(){
    return "request timeout";
}

If there is an error in the operation of the logic code here, it will also go to gets_TimeOutHandler, such as int=10/0 in logic code, will also report request timeout when accessing

Note: if the devtools code is used to control service degradation, it will take effect only after the service is restarted. If the interface invoked by the service is configured with service degradation, it needs to be consistent

Application startup class activates automatic degradation configuration

import org.springframework.cloud.client.circuitbreaker.EnableCircuitBreaker;

@EnableCircuitBreaker

It should be noted here that each configuration of service degradation will cause code expansion and redundancy, that is, if 100 api interfaces are to be degraded, 100 corresponding degradation statements need to be written, so we need to deal with service degradation in a unified manner (the general and exclusive (the above example) are separated to avoid code expansion)

@DefaultProperties(defaultFallback = "gets_TimeOutHandler")

@HystrixCommand

public String gets_TimeOutHandler(){
    return "request timeout";
}

Unified degradation processing of openFeign service call

! [insert picture description here]( https://img-blog.csdnimg.cn/1b0032df09e048ab821f2e2e9654303b.png?x-oss-process=image/watermark,type_d3F5LXplbmhlaQ,shadow_50,text_Q1NETiBA5LiA5ZCN6JC96a2E55qE56iL5bqP5ZGY,size_18,color_FFFFFF,t_70,g_se,x_16) ! [insert picture description here]( https://img-blog.csdnimg.cn/81dd0c559b31433db67e42e830ee4a72.png?x-oss-process=image/watermark,type_d3F5LXplbmhlaQ,shadow_50,text_Q1NETiBA5LiA5ZCN6JC96a2E55qE56iL5bqP5ZGY,size_18,color_FFFFFF,t_70,g_se,x_16)

Service (recovery call link)

Fuse opening parameter controller

If openfeign is used (openfeign integrates Hystrix), it can be configured in properties

#feign enable hystrix
feign.hystrix.enabled=true
#Enable httpclient
feign.httpclient.enabled=true
feign.httpclient.max-connections: 1000
#Maximum concurrent connections per service
feign.httpclient.max-connections-per-route: 500
#Request and response GZIP compression support
feign.compression.request.enabled=true
feign.compression.response.enabled=false
hystrix.threadpool.default.coreSize=500
hystrix.threadpool.default.maxQueueSize=2000
hystrix.threadpool.default.queueSizeRejectionThreshold=1500
hystrix.command.default.execution.isolation.strategy=THREAD
hystrix.command.default.circuitBreaker.requestVolumeThreshold=1500
hystrix.command.default.execution.isolation.thread.timeoutInMilliseconds=180000
hystrix.command.default.execution.timeout.enabled=true
#Sets the maximum number of threads for fallback
hystrix.command.default.fallback.isolation.semaphore.maxConcurrentRequests=500

Hystrix graphical monitoring

It is recommended to create a new service for monitoring and processing, or integrate it into the same service (it may be messy)

pom.xml dependency

<!-- hystrix-dashboard Graphical tools -->
<dependency>
    <groupId>org.springframework.cloud</groupId>
    <artifactId>spring-cloud-starter-netflix-hystrix-dashboard</artifactId>
</dependency>

Annotate the Application startup class @ EnableHystrixDashboard

import org.springframework.cloud.netflix.hystrix.dashboard.EnableHystrixDashboard;

@EnableHystrixDashboard

Start service ip: service port / hystrix

Add a dependent POM to the microservices that need to be monitored to improve the monitoring information xml

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

Add annotations on the Application startup class in the microservice to be monitored

import org.springframework.cloud.client.circuitbreaker.EnableCircuitBreaker;

@EnableCircuitBreaker

And add a Bean

/**
 * This configuration is for service monitoring and has nothing to do with the service fault tolerance itself. It is an upgrade of spring cloud
 * ServletRegistrationBean Because the default path of springboot is not / hystrix stream
 * Just configure the servlet of the context in your own project
 * @return
 */
@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;
}

The api interface in the controller performs service degradation processing (what I do here is unique degradation processing)

Start project