AbstractsComputer Science

Generic Log Analysis Tool

by Sanket Rajiv Shah

Institution: San Diego State University
Year: 2015
Record ID: 2057882
Full text PDF: http://hdl.handle.net/10211.3/137687


When you are hosting a server, developing a new chip or using any system that produces large amount of logs, you require a powerful application to analyze and debug those logs. Generally people write specialized applications that can deal with the problem of analyzing the specific format of the log they are working with. In the market there aren't applications available that are agile enough to be easily customized to work with your specific type of log file. The analysis tool is required to have multiple capabilities, different for each specific application. Fast processing speed, low memory footprint, agility, easy customizability are some of the required features. It is difficult to incorporate all of this in the applications currently available online, and make them work with dynamic user requirements. This thesis discusses the design, development, and testing of a generic log file analysis application keeping in mind the above complexity, and customizes it for its use on Oracle database log files. The application is implemented in Java and deployed as a Java Desktop Application. It uses Java Swing for the GUI and JFreeCharts for graphing purposes. Some of the features of the application include but are not limited to: 1. Visualization of interesting data using different types of charts: Line Chart, Pie Chart, Histogram 2. Performing basic statistical analysis on the data. 3. Automatic PDF report generation. 4. Analysis over multiple log files of same type (log comparison). 5. Customizable GUI environment. 6. Customizable preprocessing of data through plugin architecture. 7. Extended support for Oracle Logs. 8. Intuitive GUI design for improved post plotting navigation and archiving results.