How to rotate log files based on time rather than size in Log4j?

I use Log4j with the RollingFileAppender to create a log rotation based on size.

How can I configure it to log to each file for a certain amount of time before rotating?

For example, so that each log file contains one hour of logs, rotating at the top of each hour?

I configure Log4j programatically in Java using a Properties object (as opposed to a file)


You probably want to use a DailyRollingFileAppender. To roll them hourly, for example, you'd use a DatePattern of '.'yyyy-MM-dd-HH. For a file:


Or for your programmatic configuration:

DailyRollingFileAppender appender = new DailyRollingFileAppender();

Logger root = Logger.getRootLogger();

Unfortunately, using a DailyRollingFileAppender means that you can't limit the file size - this could be problematic if you have tons of logs in the given rolled period.



The following list shows all the date patterns which have defined by log4j,

Minutely  '.'yyyy-MM-dd-HH-mm application.log.2013-02-28-13-54
Hourly '.'yyyy-MM-dd-HH application.log.2013-02-28-13
Half-daily '.'yyyy-MM-dd-a application.log.2013-02-28-AM app.log.2013-02-28-PM
Daily '.'yyyy-MM-dd application.log.2013-02-28
Weekly '.'yyyy-ww application.log.2013-07 app.log.2013-08 
Monthly '.'yyyy-MM application.log.2013-01 app.log.2013-02

The other thing to be careful of with any rolling file appender is to make sure only one JVM access a particular log file at a time. This is because log4j caches the log file size for performance reasons, and your 'rolling' will get wonky if multiple JVMs access the same files.

Use a DailyRollingFileAppender.

In particular, setting its 'datePattern' property to '.'yyyy-MM-dd-HH would cause file to rotate every hour.

You need to use the DailyRollingFileAppender. Despite its misleading name it can be configured in a to run at configurable time periods up to minutes.

Need Your Help

NULL permitted in Primary Key - why and in which DBMS?

database database-design relational-database terminology

Further to my question "Why to use ´not null primary key´ in TSQL?"...

Getting all types that implement an interface

c# optimization reflection lambda c#-3.0

Using reflection, how can I get all types that implement an interface with C# 3.0/.NET 3.5 with the least code, and minimizing iterations?