Kafka之——扩展logback将日志输出到Kafka实例扩展

​

一、概述

本文基于《​​Kafka之——扩展logback将日志输出到Kafka实例​​​》，请先阅读《​​Kafka之——扩展logback将日志输出到Kafka实例​​》。

二、扩展

《​​Kafka之——扩展logback将日志输出到Kafka实例​​​》一文中，只能输出单个字符串到Kafka，本文对上一篇博文《​​Kafka之——扩展logback将日志输出到Kafka实例​​》做了相应的扩展，能够输出Json字符串到Kafka。主要变动如下：

1、新增JsonFormatter类

主要的作用是接收ILoggingEvent返回Json字符串，代码如下：

package com.lyz.storm.formatter.impl;

import com.lyz.storm.formatter.Formatter;

import ch.qos.logback.classic.spi.ILoggingEvent;

/**
 * Json格式的Formatter
 * @author liuyazhuang
 *
 */
public class JsonFormatter implements Formatter {
  
  private static final String QUOTE = "\"";
  private static final String COLON = ":";
  private static final String COMMA = ",";
  
  private boolean expectJson = false;
  

  @Override
  public String format(ILoggingEvent event) {
    StringBuilder sb = new StringBuilder();
    sb.append("{");
    fieldName("level", sb);
    quote(event.getLevel().levelStr, sb);
    sb.append(COMMA);
    fieldName("logger", sb);
    quote(event.getLoggerName(), sb);
    sb.append(COMMA);
    fieldName("timestamp", sb);
    sb.append(event.getTimeStamp());
    sb.append(COMMA);
    fieldName("message", sb);
    if(this.expectJson){
      sb.append(event.getFormattedMessage());
    }else{
      quote(event.getFormattedMessage(), sb);
    }
    sb.append("}");
    return sb.toString();
  }
  
  private void fieldName(String name, StringBuilder sb){
    quote(name, sb);
    sb.append(COLON);
  }
  
  private void quote(String value, StringBuilder sb){
    sb.append(QUOTE);
    sb.append(value);
    sb.append(QUOTE);
  }

  public boolean isExpectJson() {
    return expectJson;
  }

  public void setExpectJson(boolean expectJson) {
    this.expectJson = expectJson;
  }
}

2、修改logback.xml如下：

<?xml version="1.0" encoding="UTF-8"?>
<configuration>
   <appender name="KAFKA" class="com.lyz.storm.appender.KafkaAppender">
         <topic>foo</topic>
         <zookeeperHost>192.168.209.121:2181</zookeeperHost>
         <brokerList>192.168.209.121:9092</brokerList>
         <formatter class="com.lyz.storm.formatter.impl.JsonFormatter">
            <expectJson>false</expectJson>
         </formatter>
   </appender>
   <root level="debug">
         <appender-ref ref="KAFKA"/>
   </root>
</configuration>

三、测试

在Kafka服务器命令行输入如下命令：

kafka-console-consumer.sh --zookeeper localhost:2181 --topic foo --from-beginning

监听日志输出。

此时，我们运行RogueApplication类，监听到的日志如下：

[root@liuyazhuang121 local]# kafka-console-consumer.sh --zookeeper localhost:2181 --topic foo --from-beginning      
{"level":"WARN","logger":"com.lyz.storm.trend.RogueApplication","timestamp":1516674702158,"message":"This is a warning (slow state)."}
{"level":"WARN","logger":"com.lyz.storm.trend.RogueApplication","timestamp":1516674707699,"message":"This is a warning (slow state)."}
{"level":"WARN","logger":"com.lyz.storm.trend.RogueApplication","timestamp":1516674712700,"message":"This is a warning (slow state)."}

如下图所示：

四、温馨提示

大家可以到链接下载完整扩展源码