SpringBoot Log4j2发消息到Kafka以及动态添加KafkaAppender

方法一：log4j2.xml文件配置

<Appenders>

<Kafkaname ="KAFKA"topic="log4j-kafka" syncSend="false">  

  <!-- <ThresholdFilter level="error" onMatch="ACCEPT" onMismatch="DENY" />  -->

  <MarkerFiltermarker ="kafkaLog"onMatch ="ACCEPT"onMismatch= "DENY"/>  

  <PatternLayoutpattern ="%d{yyyy-MM-dd HH:mm:ss,SSS}:%4p %t (%F:%L) - %m%n"/>  

  <Property name ="bootstrap.servers" >0.0.0.0:9092</ Property> 

  <Property name ="retries" >3</ Property> 

  <Property name ="linger.ms" >1000</ Property> 

  <Property name ="buffer.memory" > 10485760</ Property> 

</Kafka>

</Appenders>

<Loggers>

   <!-- additivity属性作用是在使用当前Logger配置的Appender打印日志 -->

   <Logger name ="com.hhaip" level="INFO" additivity ="false" >

       <AppenderRefref ="CONSOLE"/>

       <AppenderRefref ="INTERPHONE"/>

       <AppenderRefref ="KAFKA"/>

   </Logger>

   <!-- Root表示所有Logger用Root中的Appender打印日志  -->

   <Root level ="INFO" >

      <AppenderRefref ="CONSOLE"/>

      <AppenderRefref ="INTERPHONE"/>

      <AppenderRefref ="KAFKA"/>

   </Root>

</Loggers>

注意不要定义key.serializer 、value.serializer和batch.size，log4j2里已经帮我们设置好了

解释：MarkerFilter是一种过滤器，marker过滤器的名称,使用时只有加上这种过滤器的日志才会输出到kafka，如下消息发送

     ThresholdFilter也是过滤器，用来指定过滤的级别,如：ERROR级别的才会发送至kafka

     PatternLayout是消息的格式

     syncSend指是否同步等待，设为false表示发送消息后立即返回，true则会等待kafka响应后返回(log4j2 2.8以上版本)

发送消息至kafka

在Java类中可以按如下方式发送消息至kafka:

log.info(Markers.KAFKA,"kafka消息{}", "log4j-" +content);

kafka收到的消息

Markers类的内容

import org.slf4j.Marker;

import org.slf4j.MarkerFactory;

publicclass Markers {

    publicstatic final Marker DB = MarkerFactory.getMarker("dbLog");

    publicstatic final Marker KAFKA = MarkerFactory.getMarker("kafkaLog");

}

方法二：动态添加（SpringBoot项目）

import java.nio.charset.Charset;

import java.util.ArrayList;

import java.util.List;

import javax.annotation.PostConstruct;

import org.apache.kafka.clients.producer.ProducerConfig;

import org.apache.logging.log4j.LogManager;

import org.apache.logging.log4j.core.Appender;

import org.apache.logging.log4j.core.Filter;

import org.apache.logging.log4j.core.Layout;

import org.apache.logging.log4j.core.Logger;

import org.apache.logging.log4j.core.LoggerContext;

import org.apache.logging.log4j.core.appender.mom.kafka.KafkaAppender;

import org.apache.logging.log4j.core.config.Configuration;

import org.apache.logging.log4j.core.config.DefaultConfiguration;

import org.apache.logging.log4j.core.config.Property;

import org.apache.logging.log4j.core.filter.MarkerFilter;

import org.apache.logging.log4j.core.layout.PatternLayout;

import org.springframework.stereotype.Component;

@Component

publicclass KafkaLogAppender {

    @PostConstruct

    publicvoid init(){

       final LoggerContext ctx = (LoggerContext) LogManager.getContext(false);

       final Configuration config = ctx .getConfiguration();

       final Logger interLogger = ctx .getLogger("com.hhaip"); //需要写日志到数据库的包名

         

        List<Property>list = new ArrayList<>();

       list .add(Property.createProperty(ProducerConfig.BOOTSTRAP_SERVERS_CONFIG ,"0.0.0.0:9092" ));

       list .add(Property.createProperty(ProducerConfig.ACKS_CONFIG , "1"));

       list .add(Property.createProperty(ProducerConfig.RETRIES_CONFIG , "3"));

       list .add(Property.createProperty(ProducerConfig.LINGER_MS_CONFIG , "10000"));

       list .add(Property.createProperty(ProducerConfig.BUFFER_MEMORY_CONFIG ,"10485760"));

       list .add(Property.createProperty(ProducerConfig.REQUEST_TIMEOUT_MS_CONFIG ,"1000"));

        Property[]props = list .toArray(new Property[list .size()]);

         

       //配置Marker过滤器(标记过滤器和方法一的一样)

        MarkerFilterfilter = MarkerFilter.createFilter("kafkaLog", Filter.Result.ACCEPT , Filter.Result.DENY);

         

        Configurationconfiguration = new DefaultConfiguration();

        Layout<String>layout = PatternLayout.createLayout("%date %message",null , configuration , null , Charset.forName ("UTF-8" ),false , false , null,null);

         

        Appenderappender = KafkaAppender.createAppender(layout,filter, "KAFKA" ,true , "log4j-kafka" , props , configuration );

         

       config .addAppender(appender );

       interLogger .addAppender(appender );

       appender .start();

       ctx .updateLoggers();   

    }

}

方法二和方法一的使用是一样的，也许要定义一个Markers类.

注意：0.0.0.0换成你自己的IP地址