当前位置：网站首页>Sentinel fusing and current limiting

Sentinel fusing and current limiting

2022-07-26 03:53:00 【RB_ VER】

summary

sentinel Take flow as the starting point , Slave flow control 、 Fusing the drop 、 Multiple dimensions such as system load protection protect the stability of services .

sentinel The main features ：
Insert picture description here
sentinel The component consists of two parts （ The front desk + backstage ）：

  • Core library （Java client ） Don't rely on any framework / library , Can run on all Java Runtime environment , At the same time Dubbo/spring cloud And other frameworks also have better support .
  • Console （dashboard） be based on spring boot Development , It can run directly after packing , No additional tomcat Etc. Application containers .

download sentinel-dashboard-1.7.0.jar, need Java8 Environment and 8080 The port is not occupied , Run the command ：

java -jar sentinel-dashboard-1.7.0.jar

Sign in sentinel Management interface http://localhost:8080, The account and password are sentinel.

Initialization project

start-up nacos8848.
newly build cloudalibaba-sentinel-service8401 modular .
pom.xml

    <dependencies>
    	<dependency>
            <groupId>org.springframework.cloud</groupId>
            <artifactId>spring-cloud-starter-openfeign</artifactId>
        </dependency>
    	<dependency>
            <groupId>com.alibaba.cloud</groupId>
            <artifactId>spring-cloud-starter-alibaba-sentinel</artifactId>
        </dependency>
    	<dependency>
            <groupId>com.alibaba.csp</groupId>
            <artifactId>sentinel-datasource-nacos</artifactId>
        </dependency>
        <dependency>
            <groupId>com.alibaba.cloud</groupId>
            <artifactId>spring-cloud-starter-alibaba-nacos-discovery</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.springframework.boot</groupId>
            <artifactId>spring-boot-devtools</artifactId>
            <scope>runtime</scope>
            <optional>true</optional>
        </dependency>
        <dependency>
            <groupId>org.projectlombok</groupId>
            <artifactId>lombok</artifactId>
            <optional>true</optional>
        </dependency>
        <dependency>
            <groupId>org.springframework.boot</groupId>
            <artifactId>spring-boot-starter-test</artifactId>
            <scope>test</scope>
        </dependency>
    </dependencies>

application.yml

server:
	port: 8401
spring:
	application:
		name: cloudalibaba-sentinel-service
	cloud: 
		nacos:
			discovery:
				server-addr: localhost:8848
		sentinel:
			transport:
				dashboard: localhost:8080 #  To configure sentinel dashboard Address 
				port: 8719 #  Default 8719 port , If occupied, then +1 scanning 

management: 
	endpoints:
		web:
			exposure:
				include: '*'

MainApp8401

@EnableDiscoveryClient
@SpringBootApplication
public class MainApp8401 {
    

	public static void main(String[] args) {
    
		SpringApplication.run(MainApp8401.class,args);
	}

}

FlowLimitController

@RestController
public class FlowLimitController {
    
	@GetMapping("/testA")
	public String testA() {
    
		return "----testA";
	}

	@GetMapping("/testB")
	public String testB() {
    
		return "----testB";
	}
}

sentinel Use lazy loading mechanism , Perform an access http://localhost:8401/testA

Insert picture description here

Flow control rules

Insert picture description here
Resource name ： Unique name , Default request path .
For source ：sentinel You can limit the flow for the caller , Fill in the microservice name , Default default（ Don't differentiate between sources ）.
Threshold type / Stand alone threshold ：

  • QPS（ Requests per second ）： When calling the API Of QPS When the threshold is reached , Carry out current limiting .
  • Number of threads ： When calling the API When the number of threads reaches the threshold , Carry out current limiting .

Cluster or not ： No need for clustering .

Flow control mode ：

  • direct ：API When the current limiting condition is reached , Direct current limiting .
  • relation ： When the associated resource reaches the threshold , Just limit yourself .
  • link ： Only the traffic on the specified link is recorded （ Specify the flow of resources from the entry resources , If the threshold is reached , Just limit the flow ）

Flow control effect ：

  • Fast failure ： Direct failure , Throw exceptions .
  • Warm Up： according to codeFactor（ Cold loading factor , Default 3） Value , From threshold /codeFactor Start , After preheating for a long time , To achieve the set QPS threshold .
  • Waiting in line ： Line up at a constant speed , Let the request pass at a constant speed , Threshold type must be set to QPS, Otherwise it will not work .

QPS Direct quick failure

Configure flow control rules ：
Insert picture description here
Insert picture description here
qps exceed 1 Then the current is limited .

relation

When associated resources /testB Of qps Threshold exceeded 1 after , Just limit the current /testA own .
Insert picture description here

preheating

Insert picture description here

The threshold for 10, Preheating duration setting 5 second , The threshold for system initialization is 10/3 About equal to 3, That is, the threshold at the beginning is 3, And then after 5 After a few seconds, the threshold slowly rises to 10.

The second kill system starts at the moment , There will be a lot of traffic coming up , It's possible to kill the system , The purpose of preheating is to protect the system , Slowly put the flow in .

Waiting in line

Insert picture description here

Line up at a constant speed , Let the request pass at an even rate , The threshold type must be set to QPS, Otherwise it will not work . The corresponding is the leaky bucket algorithm .

Insert picture description here

/testA Per second 1 Requests , If you time out, wait in line , The waiting timeout is 20000 millisecond .

Demotion rules

Insert picture description here
RT（ Mean response time , Second level ）： Requests that have an average response time that exceeds the threshold and pass through the time window >=5, When both conditions are met, degradation is triggered , Close the circuit breaker after the window period ,RT Maximum 4900（ The greater need to pass -Dcsp.sentinel.statistic.max.rt=XXXX To take effect ）.

Abnormal proportion （ Second level ）：QPS>=5 And the abnormal proportion （ Second statistics ） Threshold exceeded , Trigger downgrade ; After the time window , Turn off demotion .

Number of abnormal （ Minutes of class ）： Number of abnormal （ Minute count ） Threshold exceeded , Trigger downgrade ; After the time window , Turn off demotion .

sentinel Fuse degradation will occur when a resource in the call link is unstable （ For example, the call timeout or the exception ratio increases ）, Restrict calls to this resource , Let the request fail quickly , Avoid cascading errors that affect other resources .

When resources are demoted , Within the degradation time window of the next class , Calls to this resource are automatically blown （ The default behavior is to throw DegradeException）.

sentinel The circuit breaker is not half open .

In the half open state, the system automatically detects whether there is any abnormality in the request , If there is no abnormality, close the circuit breaker and resume use , If there is any abnormality, continue to open the circuit breaker and it is unavailable , May refer to hystrix.

RT

Insert picture description here 

@GetMapping("/testD")
public String testD() {
    
	try{
     TimeUnit.SECONDS.sleep(1); } catch(InterruptedException e) {
     e.printStackTrace(); }
	return "----testD";
}

Set up sentinel：
Insert picture description here
To configure jmeter：
Insert picture description here
Call in a second 10 Thread calls testD, And a thread needs to handle 1 second , So the circuit breaker is open , Microservices are not available .

Abnormal proportion

Insert picture description here 

@GetMapping("/testD")
public String testD() {
    
	int a = 10 / 0;
	return "----testD";
}

Insert picture description here

Number of abnormal

Insert picture description here
Insert picture description here 

@GetMapping("/testD")
public String testD() {
    
	int a = 10 / 0;
	return "----testD";
}

At the beginning of the interview, an error was reported , After five error reports , Enter fuse degradation .

hotspot key Current limiting

Hot spots are frequently accessed data . Most of the time, we want to count the most frequently visited data in a hotspot TopK data , And restrict their access . such as ：

  • goods ID Is the parameter , Count the most frequently purchased items in a period of time ID And limit .
  • user ID Is the parameter , For users who visit frequently over a period of time ID And limit .

Hot spot parameter current limiting will count the hot spot parameters in the incoming parameters , And according to the configured current limiting threshold and mode , Restrict the flow of resource calls with hotspot parameters . Hot spot parameter current limiting can be regarded as a special flow control , Only valid for resource calls with hotspot parameters .

@GetMapping("/testHotKey")
@SentinelResource(value = "testHotKey", blockHandler = "deal_testHotKey")
public String testHotKey(@RequestParam(value = "p1",require = false) String p1, @RequestParam(value = "p2",require = false) String p2) {
    
	return "testHotKey";
}

public String deal_testHotKey(String p1, String p2, BlockException ex) {
    
	return "deal_testHotKey";
}

Insert picture description here
Method testHotKey The first parameter inside is just QPS More than per second 1 Time , Downgrade immediately .

Parameter exceptions

If desired p1 Parameter when it is a special value , Its current limit is not the same as usual . Suppose that p1 The value is equal to the 5 when , Its threshold can reach 200.

Insert picture description here

System rules

The system protection rule is to control the inlet flow at the application level , From a single machine load、CPU Usage rate 、 Average RT、 entrance QPS And the number of concurrent threads , Let the system run at the maximum throughput and ensure the overall stability of the system .

System protection rules apply the whole dimension , Not the resource dimension , And only effective for inlet flow . Inlet flow is the flow into the application , such as web Service or dubbo The request received by the server , All belong to the inlet flow .

System rules support the following patterns ：

  • load The adaptive （ Only on linux/unix-like The machine works ）： Systematic load1 As an indicator of inspiration , Adaptive system protection . When the system load1 Exceed the set heuristic value , When the number of concurrent threads exceeds the estimated system capacity, system protection will be triggered （BBR Stage ）. The system capacity is determined by the maxQps * minRt It is estimated that . The setting reference value is generally CPU cores * 2.5.
  • CPU usage： When the system CPU When the utilization rate exceeds the threshold, system protection will be triggered （ Value range 0.0-1.0）, It's sensitive .
  • Average RT： When the average flow of all inlets on a single machine RT When the threshold value is reached, system protection is triggered , In milliseconds .
  • Number of concurrent threads ： When the number of concurrent threads of all the entrances on a single machine reaches the threshold, system protection will be triggered .
  • entrance QPS： When all the inlet flow on a single machine QPS When the threshold value is reached, system protection is triggered .

Insert picture description here

@SentinelResource To configure

Current limiting by resource name 

@RestController
public class RateLimitController {
    
	@GetMapping(value = "/byResource")
	@SentinelResource(value = "byResource", blockHandler = "handleException")
	public CommonResult byResource() {
    
		return new CommonResult(200,"OK",new Payment(2020L,"serial001"));
	}
	public CommonResult handleException(BlockException ex) {
    
		return new CommonResult(444,ex.getClass().getCanonicalName());
	}
}

Insert picture description here

Press URL Address restriction 

@RestController
public class RateLimitController {
    
	@GetMapping(value = "/rateLimit/byUrl")
	@SentinelResource(value = "byUrl")
	public CommonResult byUrl() {
    
		return new CommonResult(200,"OK",new Payment(2020L,"serial002"));
	}
}

Insert picture description here

Custom current limiting processing logic

RateLimitController

@RestController
public class RateLimitController {
    
	@GetMapping(value = "/rateLimit/customerBlockHandler")
	@SentinelResource(value = "customerBlockHandler", blockHandlerClass = CustomerBlockHandler.class, blockHandler = "handlerException2")
	public CommonResult customerBlockHandler() {
    
		return new CommonResult(200,"OK",new Payment(2020L,"serial002"));
	}
}

CustomerBlockHandler

public class CustomerBlockHandler{
    
	public static CommonResult handlerException1(BlockException ex) {
    
		return new CommonResult(4441,ex.getClass().getCanonicalName());
	}
	public static CommonResult handlerException2(BlockException ex) {
    
		return new CommonResult(4442,ex.getClass().getCanonicalName());
	}
}

Insert picture description here

Service function

Basic environment construction

sentinel Integrate ribbon+openFeign+fallback
Insert picture description here
Service providers
establish cloudalibaba-provider-payment9003/9004 modular 

server:
	port: 9003
spring:
	application:
		name: nacos-payment-provider
	cloud:
		nacos:
			discovery:
				server-addr: localhost:8848
			
management:
	endpoints:
		web:
			exposure:
				include: '*'

PaymentMain9003

@SpringBootApplication
@EnableDiscoveryClient
public class PaymentMain9003{
    

	public static void main(String[] args) {
    
		SpringApplication.run(PaymentMain9003.class,args);
	}

}

PaymentController

@RestController
public class PaymentController{
    
	@Value("${server.port}")
	private String serverPort;

	public static HashMap<Long,Payment> hashMap = new HashMap<>();
	static {
    
		hashMap.put(1L,new Payment(1L,"111"));
		hashMap.put(2L,new Payment(2L,"222"));
		hashMap.put(3L,new Payment(3L,"333"));
	}
	@GetMapping("/paymentSQL/{id}")
	public CommonResult<Payment> paymentSQL(@PathVariable("id") Long id) {
    
		Payment payment = hashMap.get(id);
		CommonResult<Payment> result = new CommonResult(200,"ok"+serverPort,payment);
		return result;
	} 
}

Serving consumers
establish cloudalibaba-consumer-nacos-order84 modular .

server:
	port: 84
spring:
	application:
		name: nacos-order-consumer
	cloud:
		nacos:
			discovery:
				server-addr: localhost:8848
		sentinel:
			transport:
				dashboard: localhost:8080
				port: 8719
			
service-url:
	nacos-user-service: http://nacos-payment-provider

OrderNacosMain84

@SpringBootApplication
@EnableDiscoveryClient
public class OrderNacosMain84{
    

	public static void main(String[] args) {
    
		SpringApplication.run(OrderNacosMain84.class,args);
	}

}

ApplicationContextConfig

@Configuration
public class ApplicationContextConfig{
    
	@Bean
	@LoadBalanced
	public RestTemplate getRestTemplate() {
    
		return new RestTemplate();
	}
}

CircleBreakerController

@RestController
@Slf4j
public class CircleBreakerController{
    
	public static final String SERVICE_URL = "http://nacos-payment-provider";
	
	@Resource
	private RestTemplate restTemplate;


	@RequestMapping("/consumer/fallback/{id}")
	@SentinelResource(value = "fallback")
	public CommonResult<Payment> fallback(@PathVariable("id") Long id) {
    
		CommonResult<Payment> result = restTemplate.getForObject(SERVICE_URL+"/paymentSQL/"+id,CommonResult.class,id);
		if (id == 4) {
    
			throw new IllegalArgumentException("IllegalArgumentException");
		} else if (result.getData() == null) {
    
			throw new NullPointerException("NullPointerException");
		}
		return result;
	} 
}

The service is not configured

CircleBreakerController

@RestController
@Slf4j
public class CircleBreakerController{
    
	public static final String SERVICE_URL = "http://nacos-payment-provider";
	
	@Resource
	private RestTemplate restTemplate;

	@RequestMapping("/consumer/fallback/{id}")
	@SentinelResource(value = "fallback")
	public CommonResult<Payment> fallback(@PathVariable("id") Long id) {
    
		CommonResult<Payment> result = restTemplate.getForObject(SERVICE_URL+"/paymentSQL/"+id,CommonResult.class,id);
		if (id == 4) {
    
			throw new IllegalArgumentException("IllegalArgumentException");
		} else if (result.getData() == null) {
    
			throw new NullPointerException("NullPointerException");
		}
		return result;
	} 
}

visit http://localhost:84/consumer/fallback/4 When to return to error Content .

The service only configures fallback

CircleBreakerController

@RestController
@Slf4j
public class CircleBreakerController{
    
	public static final String SERVICE_URL = "http://nacos-payment-provider";
	
	@Resource
	private RestTemplate restTemplate;

	@RequestMapping("/consumer/fallback/{id}")
	@SentinelResource(value = "fallback",fallback = "handlerFallback")
	// fallback Only responsible for business exceptions 
	public CommonResult<Payment> fallback(@PathVariable("id") Long id) {
    
		CommonResult<Payment> result = restTemplate.getForObject(SERVICE_URL+"/paymentSQL/"+id,CommonResult.class,id);
		if (id == 4) {
    
			throw new IllegalArgumentException("IllegalArgumentException");
		} else if (result.getData() == null) {
    
			throw new NullPointerException("NullPointerException");
		}
		return result;
	} 

	public CommonResult<Payment> handlerFallback(@PathVariable("id") Long id, Throwable e) {
    
		Payment payment = new Payment(id,"null");
		return new CommonResult<>("444"," The bottom of the bag is abnormal handlerFallback"+e.getMessage(),payment);
	}
}

visit http://localhost:84/consumer/fallback/4 When to return to fallback Content .

The service only configures blockHandler

CircleBreakerController

@RestController
@Slf4j
public class CircleBreakerController{
    
	public static final String SERVICE_URL = "http://nacos-payment-provider";
	
	@Resource
	private RestTemplate restTemplate;

	@RequestMapping("/consumer/fallback/{id}")
	@SentinelResource(value = "fallback", blockHandler = "blockHandler")
	// fallback Only responsible for business exceptions 
	public CommonResult<Payment> fallback(@PathVariable("id") Long id) {
    
		CommonResult<Payment> result = restTemplate.getForObject(SERVICE_URL+"/paymentSQL/"+id,CommonResult.class,id);
		if (id == 4) {
    
			throw new IllegalArgumentException("IllegalArgumentException");
		} else if (result.getData() == null) {
    
			throw new NullPointerException("NullPointerException");
		}
		return result;
	} 

	public CommonResult<Payment> blockHandler(@PathVariable("id") Long id,  BlockException e) {
    
		Payment payment = new Payment(id,"null");
		return new CommonResult<>("444","BlockException "+e.getMessage(),payment);
	}
}

Insert picture description here
One visit per second http://localhost:84/consumer/fallback/4 When to return to error Content .
Multiple visits per second http://localhost:84/consumer/fallback/4 When to return to blockHandler Content .

Service fuse configuration fallback and blockHandler

CircleBreakerController

@RestController
@Slf4j
public class CircleBreakerController{
    
	public static final String SERVICE_URL = "http://nacos-payment-provider";
	
	@Resource
	private RestTemplate restTemplate;

	@RequestMapping("/consumer/fallback/{id}")
	@SentinelResource(value = "fallback", fallback = "handlerFallback", blockHandler = "blockHandler")
	// fallback Only responsible for business exceptions 
	public CommonResult<Payment> fallback(@PathVariable("id") Long id) {
    
		CommonResult<Payment> result = restTemplate.getForObject(SERVICE_URL+"/paymentSQL/"+id,CommonResult.class,id);
		if (id == 4) {
    
			throw new IllegalArgumentException("IllegalArgumentException");
		} else if (result.getData() == null) {
    
			throw new NullPointerException("NullPointerException");
		}
		return result;
	} 

	public CommonResult<Payment> blockHandler(@PathVariable("id") Long id,  BlockException e) {
    
		Payment payment = new Payment(id,"null");
		return new CommonResult<>("444","BlockException "+e.getMessage(),payment);
	}
	public CommonResult<Payment> handlerFallback(@PathVariable("id") Long id, Throwable e) {
    
		Payment payment = new Payment(id,"null");
		return new CommonResult<>("444"," The bottom of the bag is abnormal handlerFallback"+e.getMessage(),payment);
	}	
}

Insert picture description here

One visit per second http://localhost:84/consumer/fallback/1 Return normal data .
Multiple visits per second http://localhost:84/consumer/fallback/1 When to return to blockHandler Content .
One visit per second http://localhost:84/consumer/fallback/4 When to return to fallback Content .
Multiple visits per second http://localhost:84/consumer/fallback/4 When to return to blockHandler Content .

Service fuse configuration exceptionsToIgnore

CircleBreakerController

@RestController
@Slf4j
public class CircleBreakerController{
    
	public static final String SERVICE_URL = "http://nacos-payment-provider";
	
	@Resource
	private RestTemplate restTemplate;

	@RequestMapping("/consumer/fallback/{id}")
	@SentinelResource(value = "fallback", fallback = "handlerFallback", blockHandler = "blockHandler", exceptionsToIgnore = {
    IllegalArgumentException.class})
	// fallback Only responsible for business exceptions 
	public CommonResult<Payment> fallback(@PathVariable("id") Long id) {
    
		CommonResult<Payment> result = restTemplate.getForObject(SERVICE_URL+"/paymentSQL/"+id,CommonResult.class,id);
		if (id == 4) {
    
			throw new IllegalArgumentException("IllegalArgumentException");
		} else if (result.getData() == null) {
    
			throw new NullPointerException("NullPointerException");
		}
		return result;
	} 

	public CommonResult<Payment> blockHandler(@PathVariable("id") Long id,  BlockException e) {
    
		Payment payment = new Payment(id,"null");
		return new CommonResult<>("444","BlockException "+e.getMessage(),payment);
	}
	public CommonResult<Payment> handlerFallback(@PathVariable("id") Long id, Throwable e) {
    
		Payment payment = new Payment(id,"null");
		return new CommonResult<>("444"," The bottom of the bag is abnormal handlerFallback"+e.getMessage(),payment);
	}	
}

If the IllegalArgumentException, No more fallback Methods , No degradation effect .

Service fuse configuration OpenFeign

modify cloudalibaba-consumer-nacos-order84 modular
pom.xml

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

application.yml

#  Activate sentinel Yes feign Support for 
feign:
	sentinel:
		enable: true

OrderNacosMain84

@SpringBootApplication
@EnableDiscoveryClient
@EnableFeignClients
public class OrderNacosMain84{
    

	public static void main(String[] args) {
    
		SpringApplication.run(OrderNacosMain84.class,args);
	}

}

PaymentService

@FeignClient(value = "nacos-payment-provider",fallback = PaymentFallbackService.class)
public Interface PaymentService {
    
	@GetMapping(value = "/paymentSQL/{id}")
	public CommonResult<Payment> paymentSQL(@PathVariable("id") Long id);
}

PaymentFallbackService

@Component
public class PaymentFallbackService implements PaymentService {
    
	@Override
	public CommonResult<Payment> paymentSQL(Long id) {
    
		return new CommonResult<>("444"," Service degradation returns ",new Payment(id,"null"));
	}
}

CircleBreakerController

@RestController
@Slf4j
public class CircleBreakerController{
    
	public static final String SERVICE_URL = "http://nacos-payment-provider";
	
	@Resource
	private RestTemplate restTemplate;
	@Resource
	private PaymentService paymentService;
	@GetMapping(value = "/consumer/paymentSQL/{id}")
	public CommonResult<Payment> paymentSQL(@PathVariable("id") Long id) {
    
		return paymentService.paymentSQL(id);
	}
	
}

Rule persistence

Once the application is restarted ,sentinel The rules will disappear , The production environment needs to persist the configuration rules .

Persistence of current limiting configuration rules into Nacos preservation , Just refresh 8401 Some rest Address ,sentinel You can see the flow control rules of the console , as long as nacos The configuration is not deleted , in the light of 8401 On sentinel The flow control rules on continue to take effect .

modify cloudalibaba-sentinel-service8401 modular .
pom.xml

<dependency>
	<groupId>com.alibaba.csp</groupId>
	<artifactId>sentinel-datasource-nacos</artifactId>
</dependency>

application.yml

#  Activate sentinel Yes feign Support for 
spring:
	cloud:
		sentinel:
			datasource:
				dsl:
					nacos: 
						server-addr: localhost:8848
						dataId: cloudalibaba-sentinel-service
						groupId: DEFAULT_GROUP
						data-type: json
						rule-type: flow

Insert picture description here
resource： Resource name
limitApp： Source application
grade： Threshold type ,0 Number of threads ,1 Express qps
count： Stand alone threshold
strategy： Flow control mode ,0 It means direct ,1 Express Association ,2 Represents a link
controlBehavior： Flow control effect ,0 Express a quick failure ,1 Express warm up,2 Stand in line
clusterMode： Cluster or not

start-up 8401, Refresh http://localhost:8401/rateLimit/byUrl
Insert picture description here

原网站

版权声明
本文为[RB_ VER]所创，转载请带上原文链接，感谢
https://yzsam.com/2022/201/202207182208282345.html

边栏推荐

猜你喜欢

随机推荐