Perf4J is a set of utilities for calculating and displaying performance statistics for Java code. For developers who are familiar with logging frameworks such as log4j or logback, an analogy helps to describe Perf4J:
Perf4J is to System.currentTimeMillis() as log4j is to System.out.println()
How is this relevant to Perf4J? Consider that before good logging frameworks were widely available, developers new to Java would often print debugging statements using System.out.println(). Later they would discover they wanted log statements to go to a dedicated file, perhaps a log file that was rolled daily. Then they would discover they wanted finer control of which log statements would be written, and they wanted it to be possible to only have some log statements written in some environments without changing any code. Thus, the full set of features that log4j provides comes out of the original desire to have a "better" System.out.println() for logging statements.
Similarly, when new Java developers discover that they need to time specified blocks of code for performance logging and monitoring reasons, they often do something like this:
long start = System.currentTimeMillis();
System.out.println("ms for block n was:"+(System.currentTimeMillis() - start));
Only caliculating time of execution no not only that have option to create graphs, such as aggregated performance statistics like mean, minimum, maximum, standard deviation and transactions per second over a set time span. They would like graphs of these values in real time to detect problems on a running server, or they would like to expose performance metrics through JMX. In addition, they want their timing statements to work with the common logging frameworks like log4j (and, more recently, SLF4J).