Application Performance Statistics

• Layer Wise Statistics: Performance summary gives layer wise break up of the performance statistics. Each layer can be examined in detail to see API times within it. InfraRED comes packaged with default layer definitions bundles default layer definitions such as Web, Session (Business Logic) and JDBC. These definitions can be customized on need basis by writing simple aspects.
• Average Operation and API Timing: Average time taken by each API can be seen. APIs taking longer than a specified threshold are shown in a different color. First, last, min, max execution times are also shown.
• Call Sequence: The call sequence of the operation can be seen as an execution tree.
• Http Request/Response Summary: The total number of hits to the server and the average response timings are captured.
• Last Invocation: Performance statistics for the last n requests made to the application are shown under Last Invocation tab. This can be very useful to debug performance problems of the most recent hits to the application.

• SQL Query View: The metrics grouped under JDBC statistics is provided as a single page vital information.
• Summary of the JDBC APIs invoked by the application shows,
  • Top n expensive SQL queries. Helps in determining which queries need tuning.
  • Top n most frequently executed SQL queries. Any small improvement made to these queries will have a bigger impact on the overall performance of the application.

• Ant script to automatically instrument J2EE Apps
• Ear and War format supported on
  • Weblogic
  • Jboss
  • Tomcat
• Supports AspectWerkz, load-time weaving (both in Weblogic and JBoss). XML Based pointcut definition, which is easier to create and change, is available.

• Detailed Call trace: For each method InfraRED provides a detailed call trace showing all the methods called from within and the time spent in each of them as a proportion of the method execution time. Average time taken by the individual APIs is shown in the call trace. JDBC statistics (Time taken by the JDBC APIs, number of calls made to JDBC APIs and SQL queries) is shown separately for the selected top-level APIs.
• Exclusive and Inclusive time: API times are split into two components, time spent only in executing the API (exclusive time) and total time including time spent in the child APIs (inclusive time). By sorting on exclusive time, the bottlenecks in an application can be determined with greater ease.
• Click through access to the statistics of remote APIs.

• Web based user interface, which give detailed performance statistics.
• Multiple application instances can be integrated with InfraRED. Performance statistics for all the instances can be viewed individually or in aggregation.
• All performance metrics mentioned earlier are easily accessible via InfraRED web GUI.
• Performance metrics for multiple application instances can be seen on InfraRED Web GUI. Two or more application instances can be chosen and their statistics are presented one after the other. On clicking on remote calls made from one application instance to the other application instance, user is navigated to the call tree of second application.

• Local/Remote Collection of Performance data : InfraRED provides an ability to collect the performance statistics for an application on a remotely located infrared server. This feature becomes handy when the application is deployed on a cluster of servers or when there is a need to monitor a system that is composed of multiple applications. The data can be collected in a central server and can be viewed through the InfraRED Web GUI either individually for a specific application or a server or in an aggregated way.

• Provides layer wise break up of the execution time.
  • Provides summary information for a particular layer.
  • Layer details - Shows a list of methods belonging to a particular layer along with other metrics like total time, average time, min, max time etc.

• The developer need not edit any of his/her source files to monitor application with InfraRED.
• InfraRED has been designed and tuned to have very little overhead on the application.
• Due to low overhead, it can be used in production environments.

• Summary of performance statistics can be exported into Excel spreadsheets.
• Snapshots of the current session can be saved and reloaded later for analysis.