Shulou(Shulou.com)06/02 Report--

Then in the previous article, the needle made some summary of the common problems of Spring Cloud again.

1. The first request failed after integrating Hystrix

1.1 cause analysis

The default timeout for Hystrix is 1 second, and if you don't get a response within 1 second, you will enter the fallback logic.

To understand the springcloud architecture, please add: 3536247259, due to the lazy loading mechanism of Spring.

The first request tends to be slow, so on some machines (especially those with low configuration), the first request may take more than 1 second.

1.2 solution

There are many ways to solve this problem. Here are some relatively simple solutions:

1) method 1: configure hunger loading for Ribbon.

Ribbon: eager-load: enabled: true clients: client1,client2

For Zuul:

Zuul: ribbon: eager-load: enabled: true

2) method 2: extend the timeout of Hystrix. The example is as follows

Hystrix.command.default.execution.isolation.thread.timeoutInMilliseconds:5000

This configuration changes the timeout of Hystrix to 5 seconds.

3) method 3: disable Hystrix timeout, as shown in the following example

Hystrix.command.default.execution.timeout.enabled: false

Method 4: for Feign, you can also disable Hystrix for Feign. The example is as follows

Feign.hystrix.enabled: false

This enables Hystrix support to be disabled globally for Feign. However, this method is extreme and is generally not recommended.

2. Turbine aggregate data is incomplete

This problem occurs with Turbine in some versions of Spring Cloud, such as Brixton SR5. The direct view of this problem is as follows:

Turbine is used to aggregate multiple micro-services, but only some of the monitoring data of micro-services can be seen on Hystrix Dashboard.

Phenomenon description:

For example, the configuration of Turbine is as follows:

Turbine is supposed to aggregate the monitoring data of the two micro-services cloud-consumer-movie,cloud-consumer-movie-feign-hystrix-fallback-stream. However, when you open Hystrix Dashboard, you will find that only part of the monitoring data of micro-services are displayed on Dashboard.

Solution:

This problem occurs when Turbine-aggregated microservices are deployed on the same host.

Solution one:

Configure different hostname for each microservice, and set preferIpAddress to false or not.

Eureka: client: serviceUrl: defaultZone: http://127.0.0.1:8001/eureka/ instance: hostname:ribbon # configure hostname

Solution 2:

Set turbine.combine-host-port = true

Turbine: appConfig: cloud-consumer-movie,cloud-consumer-movie-feign-hystrix-fallback-stream clusterNameExpression: "'default'" combine-host-port:true

Method 3:

Upgrade Spring Cloud to Camden or later. Of course, you can also upgrade Spring Cloud Netflix to the latest stable version above 1.2.0 separately (it is generally not recommended to upgrade Spring Cloud Netflix separately, as it may conflict with other components of Spring Cloud).

This is because the default value for turbine.combine-host-port in older versions is false. Spring Cloud is aware of the problem, so the default value for this property is set to true in the new version. The solution and method are essentially the same.

Related code

Org.springframework.cloud.netflix.turbine.TurbineProperties.combine-HostPort org.springframework.cloud.netflix.turbine.CommonsInstanceDiscovery.getInstance (String, Boolean)

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.