SLF4J and LOGBack

Next Generation Logging!

31-08-2010

LOGBack : History

It’s a successor of log4j

LOGBack together with SLF4J designed to be the next generation logging framework

There is no active developer community for log4j

What is SLF4J?

Remember commons-logging?SLF4J is a façade for different logging toolsDifferent logging tools can be plugged without any code changeNo classloader and memory leak issueBetter support of MDC (Later!)No difference using with LOGBack but advantageous on the way!

Hello SLF4J

import org.slf4j.Logger;import org.slf4j.LoggerFactory;

public class HelloSLF4J { private static final Logger logger = LoggerFactory.getLogger(HelloSLF4J.class); public static void main(String[] args) {

logger.info("Hello World"); }}

logger.debug("The entry is {}.", entry);

SLF4J Enforcing Best Practices!

if(logger.isDebugEnabled()) {logger.debug("Entry number: " + i + " is " + String.valueOf(entry[i]));

}

•Parameterised Logging•String manipulation after checking loglevel

Exception Logging

logger.error("some accompanying message", e);

try {  Integer i = Integer.valueOf(s);} catch (NumberFormatException e){  logger.error("Failed to format {}", s, e);}

What is LOGBack?

“LOGBack is faster, reliable and feature upgraded logging tool compared to its predecessors“

SLF4J and Logback : Dream Dates!

Logback implements SLF4J natively

No computational and memory overhead

Fully compliant to SLF4J

Faster they say! 10 times!! Smaller Blue Print

Core, classic and access modules

Pre-Requirements

logback.xml/logback.groovy

slf4j-api.jar

logback-classic.jar

logback-core.jar

Why Logback

Automatic Reloading• LOGLEVELs can be changed on the fly. • <configuration scan="true" scanPeriod="30 seconds" >

Prudent Model Recovery (3 times slower)• Multiple JVM can share same log file.• Graceful recovery from IO failures(100% disk usage) without

process restart. File Server recovery will let log files working automatically

Filters And MDCRolling Policies• Compress and Roll• Housekeeping with maxHistory property

Why LogbackConditional Processing• Good For Dev/Prod switch

Is It Worth A Use?

Why Logback

Stack Traces Pointing jar files

logback.xmlConsoleAppenderConsoleAppender

RollingFileAppenderRollingFileAppender

rollover daily or whenever the file size reaches 100MB rollover daily or whenever the file size reaches 100MB

logback.groovy

import ch.qos.logback.classic.encoder.PatternLayoutEncoderimport ch.qos.logback.core.ConsoleAppender

import static ch.qos.logback.classic.Level.DEBUG

appender("STDOUT", ConsoleAppender) { encoder(PatternLayoutEncoder) { pattern = "%d{HH:mm:ss.SSS} [%thread] %-5level %logger{36} -

%msg%n" }}root(DEBUG, ["STDOUT"])

Best Practice

Though the samples here used are xml for our understanding, groovy is simpler

Maven will take care most!• src/test/resources/logback-test.xml

• src/main/resources/logback.xml

• No setting required, scanner checks groovy, test and production in an order!

MDC

MDC : Map Diagnostic Context• Helps to uniquely stamp requests

• Similar to NDC : Nested Diagnostic Context

• Correlating the logs effectively

The MDC manages contextual information on a per thread basis(and its children)

MDC Use Case: User Based Loggingpublic class UserServletFilter implements Filter { private final String USER_KEY = "username"; public void doFilter(ServletRequest request, ServletResponse response, FilterChain chain) throws IOException, ServletException { HttpServletRequest req = (HttpServletRequest) request; Principal principal = req.getUserPrincipal(); if (principal != null) { String username = principal.getName(); MDC.put(USER_KEY, username); } try { chain.doFilter(request, response); } finally { MDC.remove(USER_KEY); } }}

MDC Use Case: User Based Logging

  <appender name="CONSOLE“ class="ch.qos.logback.core.ConsoleAppender">    <layout class="ch.qos.logback.classic.PatternLayout">      <Pattern>%-4r [%thread] %-5level C:%X{username} - %msg%n</Pattern>    </layout>   </appender>  

MDC Use Case: InsertingServletFilter

%X{req.remoteHost}

%X{req.requestURI}%n%d - %m%n

MDC Use Case: SiftingAppender

Separate logging based runtime attributeslogger.debug("Application started");

MDC.put("userid", "Alice");

logger.debug("Alice says hello");

Filters: LevelFilter

<appender name="CONSOLE" class="ch.qos.logback.core.ConsoleAppender">    <filter class="ch.qos.logback.classic.filter.LevelFilter">      <level>INFO</level>      <onMatch>ACCEPT</onMatch>      <onMismatch>DENY</onMismatch>    </filter>    <encoder>      <pattern>        %-4relative [%thread] %-5level %logger{30} - %msg%n      </pattern>    </encoder>  </appender>

Filters: ThresholdFilter

<configuration>  <appender name="CONSOLE"    class="ch.qos.logback.core.ConsoleAppender">    <!-- deny all events with a level below INFO, that is TRACE and DEBUG -->    <filter class="ch.qos.logback.classic.filter.ThresholdFilter">      <level>INFO</level>    </filter>    <encoder>      <pattern>        %-4relative [%thread] %-5level %logger{30} - %msg%n      </pattern>    </encoder>  </appender>  <root level="DEBUG">    <appender-ref ref="CONSOLE" />  </root></configuration>

Filters:EvaluatorFilter

<configuration>  <appender name="STDOUT" class="ch.qos.logback.core.ConsoleAppender">    <filter class="ch.qos.logback.core.filter.EvaluatorFilter">            <evaluator>

        <expression>message.contains("billing")</expression>      </evaluator>      <OnMismatch>NEUTRAL</OnMismatch>      <OnMatch>DENY</OnMatch>    </filter>    <encoder>      <pattern>        %-4relative [%thread] %-5level %logger - %msg%n      </pattern>    </encoder>  </appender>

  <root level="INFO">    <appender-ref ref="STDOUT" />  </root></configuration>

Good To Know!

Logback-access• powerful HTTP-access log with tomcat and jetty

• UI based

• JMX support

• Statistical Data of UI accesses

• TeeFilter, CountingFilter, ViewStatusMessagesServlet

Tools : SLF4J Migrator tool

DANKE!!!