Prerequisites

1. Project migrated to gradle;

2. Using Log4J2 as logging library;

Steps

1. Add Spring Cloud Sleuch dependency to project.
   In build.gradle, add dependency as below.

   dependencies {
       compile 'org.springframework.cloud:spring-cloud-starter-sleuth:2.2.6.RELEASE'
   }

2. Deploy project to target server.

3. Add Socket appender to write log via Socket call.
   In log4j2.xml, add socket appender and enable it.

   <Configuration status="WARN" monitorInterval="30">
       <Properties>
           <Property name="LOG_PATTERN">%d{yyyy-MM-dd HH:mm:ss.SSS} %-5level --- [%t] %notEmpty{ [%X{x-apigw-requestId}]} -  %logger{36} - %msg%n</Property>
           <Property name="LOG_FILE_NAME">user-management</Property>
           <Property name="LOG_PATH">c:\log\user-management</Property>
       </Properties>
       <Appenders>
           <Socket name="socket" host="10.3.11.114" port="19722" protocol="TCP">
               <JsonLayout properties="true" compact="true" eventEol="true" />
               <PatternLayout pattern="${LOG_PATTERN}" />
           </Socket>
           <Async name="async">
               <AppenderRef ref="socket" />
           </Async>
       </Appenders>
       <Loggers>
           <root level="info">
               <AppenderRef ref="async" />
           </root>
       </Loggers>
   </Configuration>

4. Create Logstash pipeline.
   In logstash server, create config file in /etc/logstash/conf.d/ and add content below.
   Command

   ## Sample service name: User management service.
   sudo vi /etc/logstash/conf.d/user-management-service-logstash.conf

   Content

   ## This sample use port 19721 to collect log from socket call in JSON format.
   input {
     tcp {
         port => 19721
         codec => json
     }
     stdin {}
   }
   
   output {
     stdout {
       codec => rubydebug
     }
     elasticsearch {
       action => "index"
       hosts => [ "10.3.11.114:9200" ]
       index => "user-management-service-%{+YYYY.MM.dd}"
     }
   }

5. Add index pattern in Kibana.
   In ELK main menu, click Stack Management > Index Patterns > Create index pattern.

   

6. Configure new index pattern
   In Index pattern name, input user-management-service* and un-select include system and hidden indices. After click Net step, select @timestamp in Time field and click Create index pattern.

   
   

7. Add index pattern to log.
   In ELK main screen, click Logs > Settings
   In Log indices, append index pattern created in previous step.

   

8. Test.
   In ELK main screen, click Log and select log in filter, if result shown, means integration success.

   

However, it only apply for controller to diagnosis. spin and trace id will miss when exception thrown. This article will status how to resolve this problem.

In Springboot MVC project, it simply add new ControllerAdvice to capture exception and it can be resolve.

import org.apache.logging.log4j.LogManager;
import org.apache.logging.log4j.Logger;
import org.springframework.http.HttpStatus;
import org.springframework.web.bind.annotation.ControllerAdvice;
import org.springframework.web.bind.annotation.ExceptionHandler;
import org.springframework.web.bind.annotation.ResponseBody;
import org.springframework.web.bind.annotation.ResponseStatus;

@ControllerAdvice
public class ControllerExceptionHandler {
    private static final Logger logger= LogManager.getLogger(ControllerExceptionHandler.class);

    @ResponseStatus(value = HttpStatus.INTERNAL_SERVER_ERROR)
    @ExceptionHandler(Exception.class)
    @ResponseBody
    public void handleInternalError(Exception exception) {
        logger.error(exception.getMessage(),exception);
    }
}

Screenshot of test result, trace Id and spin Id shown in log.