Log (common log framework)

Common log framework

Log4j

Log4j yes Apache One of the Java The logstore of , By using Log4j, We can control the destination of log information transmission （ Console 、 file 、 Database etc. ）; We can also control the output format of each log ; By defining the level of each log message , We can control the log generation process more carefully .

Logback

Logback, One “ reliable 、 Universal 、 Fast and flexible Java Log framework ”.logback Currently divided into three modules ：logback-core,logback- classic and logback-access.logback-core It is the basic module of the other two modules .logback-classic yes log4j An improved version of . Besides logback-classic Complete implementation SLF4J API So that you can easily change to other log system , Such as log4j or JDK14 Logging.logback-access Module and Servlet Containers （ Such as Tomcat and Jetty） Integrate , In order to provide HTTP Access logging . Please note that , You can go to logback-core It's easy to build your own modules .

Log4j2

Apache Log4j 2 It's right Log4j The upgrade , It's better than its predecessor Log4j 1.x Provides significant improvements , And provides Logback Many of the improvements available in , And fixed Logback Some problems in Architecture .

Now the best Java The log framework is Log4j2, Not one of them. . According to official tests , In multithreaded environment ,Log4j2 Asynchronous logging is better . In the asynchronous log ,Log4j2 Use independent threads to execute I/O operation , Can greatly improve the performance of the application .

Performance comparison of the three frameworks ：

effect

debugging

stay Java During project commissioning , Viewing the stack information can easily know the running state of the current program , The output log is convenient for recording the previous running results of the program . If you use a lot of System.out perhaps System.err, This is the most convenient and effective way , But it's not professional enough .

The wrong location

Don't think that if the project can run correctly, you can rest easy , After the project has been running for some time , Maybe due to data problems , Network problems , Memory problems and other exceptions . At this time, logs can help developers or operation and maintenance personnel quickly locate the wrong location , Propose solutions .

Data analysis

The rise of big data , Make a lot of log analysis possible ,ELK It also lowers the threshold of log analysis . The log contains a lot of user data , Including click behavior , Interest preference, etc , The user portrait can guide the company's next strategic direction .

Log4j2

Apache Log4j 2 It's right Log4j The upgrade , It's better than its predecessor Log4j 1.x Provides significant improvements , And provides Logback Many of the improvements available in , And fixed Logback Some problems in Architecture . It's the best Java Log framework , Not one of them. .

features

API Separate

Log4j Of API Separate from implementation , So that application developers can clearly understand which classes and methods can be used , While ensuring forward compatibility . This makes Log4j Teams can improve in a safe and compatible way .

Performance improvement

Log4j 2 Include based on LMAX Disruptor Library's next generation of asynchronous recorders . In a multithreaded scenario , The throughput of asynchronous recorder is better than Log4j 1.x and Logback high 18 times , Low latency .

Auto reload configuration

And Logback equally ,Log4j 2 You can automatically reload its configuration when you modify it . And Logback Different , It will not lose log events when reconfiguration occurs .

Advanced filtering

And Logback equally ,Log4j 2 Support based on Log Context data in events , Mark , Regular expressions and other components for filtering . Besides , Filters can also be associated with recorders . And Logback Different , You can use generic... In any of these cases Filter class .

Plug-in architecture

Log4j Use plug-in mode to configure components . therefore , You don't have to write code to create and configure Appender,Layout,Pattern Converter etc. . When configured ,Log4j Automatically identify plug-ins and use them .

No garbage mechanism

During steady state logging ,Log4j 2 Garbage free in stand-alone applications , stay Web Low garbage in applications . This reduces the pressure on the garbage collector , And it can provide better response performance .

Use

Citation depends on

it is to be noted that , Simultaneous introduction spring-boot-starter and spring-boot-starter-web When you depend on , Need to be in spring-boot-starter Exclude from spring-boot-starter-logging rely on , because spring-boot-starter-web Depend on spring-boot-starter, according to Maven Dependency relationship , stay spring-boot-starter Exclude from spring-boot-starter-logging Success is ruled out by dependence .

<dependency>
    <groupId>org.springframework.boot</groupId>
    <artifactId>spring-boot-starter-web</artifactId>
    <!--  exclude spring boot The default log logback -->
    <exclusions>
        <exclusion>
            <groupId>org.springframework.boot</groupId>
            <artifactId>spring-boot-starter-logging</artifactId>
        </exclusion>
    </exclusions>
</dependency>

<!--  introduce log4j2 rely on  -->
<dependency>
    <groupId>org.springframework.boot</groupId>
    <artifactId>spring-boot-starter-log4j2</artifactId>
</dependency>

Add configuration file

By default ,Log4j2 stay classpath The following search name is log4j2.xml Configuration file for . You can also use Java The start command specifies the full path of the configuration file .-Dlog4j.configurationFile=opt/demo/log4j2.xml, You can still use it Java The code specifies the configuration file path ：

public static void main(String[] args) {
    

    LoggerContext loggerContext = (LoggerContext) LogManager.getContext(false);
    File file = new File("opt/demo/log4j2.xml");
    loggerContext.setConfigLocation(file.toURI());

}

Asynchronous profile

Log4j-2.9 And later versions require disruptor-3.3.4.jar Or later . stay Log4j-2.9 Before , need disruptor-3.0.0.jar Or later

By default , Asynchronous recorders do not pass positions to I / O Threads . If you need location information for a layout or custom filter , You need to set... In the configuration of all related recorders “includeLocation = true”, Including the root logger .

1. introduce disruptor rely on

   <dependency>
       <groupId>com.lmax</groupId>
       <artifactId>disruptor</artifactId>
       <version>3.4.2</version>
   </dependency>
   

2. stay src/java/resources Catalog add log4j2.component.properties The configuration file

   #  Set asynchronous logging system properties 
   log4j2.contextSelector=org.apache.logging.log4j.core.async.AsyncLoggerContextSelector
   

3. The configuration is as follows ：

   adopt log.info(“ Whether it is an asynchronous log ：{}”, AsyncLoggerContextSelector.isSelected()); You can check whether it is an asynchronous log .

   <?xml version="1.0" encoding="UTF-8"?>
   <Configuration>
   
       <Properties>
           <!--  Log output level  -->
           <Property name="LOG_INFO_LEVEL" value="info"/>
           <!-- error The level of log  -->
           <Property name="LOG_ERROR_LEVEL" value="error"/>
           <!--  Create a file in the current directory called log Directory is the directory where logs are stored  -->
           <Property name="LOG_HOME" value="./log"/>
           <!--  File log storage directory  -->
           <Property name="LOG_ARCHIVE" value="./log/archive"/>
           <!--  Module name ,  Affect log configuration name , Log file name , Configure according to your own project  -->
           <Property name="LOG_MODULE_NAME" value="spring-boot"/>
           <!--  Log file size , Beyond this size will be compressed  -->
           <Property name="LOG_MAX_SIZE" value="100 MB"/>
           <!--  How many days of logs to keep  -->
           <Property name="LOG_DAYS" value="15"/>
           <!-- Format of output log ：%d Indicates the date ,%thread Represents the thread name ,%-5level： The level is shown from the left 5 Character width , %msg： Log message ,%n Is a newline  -->
           <Property name="LOG_PATTERN" value="%d [%t] %-5level %logger{0} - %msg%n"/>
           <!--interval Property is used to specify how often to scroll -->
           <Property name="TIME_BASED_INTERVAL" value="1"/>
       </Properties>
   
       <Appenders>
           <!--  Console output  -->
           <Console name="STDOUT" target="SYSTEM_OUT">
               <!-- Format of output log -->
               <PatternLayout pattern="${LOG_PATTERN}"/>
               <!-- Console output only level And above （onMatch）, Other direct rejection （onMismatch）-->
               <ThresholdFilter level="${LOG_INFO_LEVEL}" onMatch="ACCEPT" onMismatch="DENY"/>
           </Console>
   
           <!--  This will print out all the info Above grade ,error Log below level , Each time the size exceeds size Or satisfied TimeBasedTriggeringPolicy, The log will be automatically saved under the folder created by the date of year and compressed , As archive -->
           <!-- Asynchronous logs are automatically refreshed in batches , So will immediateFlush Property is set to false-->
           <RollingRandomAccessFile name="RollingRandomAccessFileInfo" fileName="${LOG_HOME}/${LOG_MODULE_NAME}-infoLog.log" filePattern="${LOG_ARCHIVE}/${LOG_MODULE_NAME}-infoLog-%d{yyyy-MM-dd}-%i.log.gz" immediateFlush="false">
               <Filters>
                   <!-- If it is error Level reject , Set up  onMismatch="NEUTRAL"  You can let the log go through subsequent filters -->
                   <ThresholdFilter level="${LOG_ERROR_LEVEL}" onMatch="DENY" onMismatch="NEUTRAL"/>
                   <!-- If it is info\warn Output -->
                   <ThresholdFilter level="${LOG_INFO_LEVEL}" onMatch="ACCEPT" onMismatch="DENY"/>
               </Filters>
               <PatternLayout pattern="${LOG_PATTERN}"/>
               <Policies>
                   <!--interval Property is used to specify how often to scroll , Based on the current filePattern Set yes 1 The day rolls once -->
                   <TimeBasedTriggeringPolicy interval="${TIME_BASED_INTERVAL}"/>
                   <SizeBasedTriggeringPolicy size="${LOG_MAX_SIZE}"/>
               </Policies>
               <!-- DefaultRolloverStrategy Property if not set , By default, it can save at most in the same folder 7 File -->
               <DefaultRolloverStrategy max="${LOG_DAYS}"/>
           </RollingRandomAccessFile>
   
           <!-- Record only error Log above level , And info Level logs are saved in different files -->
           <RollingRandomAccessFile name="RollingRandomAccessFileError" fileName="${LOG_HOME}/${LOG_MODULE_NAME}-errorLog.log" filePattern="${LOG_ARCHIVE}/${LOG_MODULE_NAME}-errorLog-%d{yyyy-MM-dd}-%i.log.gz" immediateFlush="false">
               <Filters>
                   <ThresholdFilter level="${LOG_ERROR_LEVEL}" onMatch="ACCEPT" onMismatch="DENY"/>
               </Filters>
               <PatternLayout pattern="${LOG_PATTERN}"/>
               <Policies>
                   <TimeBasedTriggeringPolicy interval="${TIME_BASED_INTERVAL}"/>
                   <SizeBasedTriggeringPolicy size="${LOG_MAX_SIZE}"/>
               </Policies>
               <DefaultRolloverStrategy max="${LOG_DAYS}"/>
           </RollingRandomAccessFile>
   
       </Appenders>
   
       <Loggers>
           <!--  Development environment use  -->
           <!--<Root level="${LOG_INFO_LEVEL}"> <AppenderRef ref="STDOUT"/> </Root>-->
   
           <!--  test , Production environment use  -->
           <Root level="${LOG_INFO_LEVEL}" includeLocation="false">
               <AppenderRef ref="RollingRandomAccessFileInfo"/>
               <AppenderRef ref="RollingRandomAccessFileError"/>
           </Root>
       </Loggers>
   
   </Configuration>
   

​