In-depth analysis and diagnostics of application issues using Itcam for WebSphere

Introduction

The current identification, isolation, diagnosis and repair of Java EE application problems is a very troublesome thing, although before the application on the line, we have undergone rigorous functional and performance testing, and through strict coding standards to ensure the quality of applications, but often encountered on the line after a period of time, The application system is always in an unstable state, and then after a period of slow analysis to find some problems in the application development, and finally fix it. In addition, most applications in their life cycle will have some problems, and most of these are occasional events, often make maintenance and developers headaches, often consumes some of the cost of users to ask technical experts to solve, time-consuming, consumption, and knowledge can not be effectively accumulated.

Below, we will use the IBM Tivoli Composite Application Manager for WebSphere (hereinafter referred to as ITCAM for WebSphere) to explain how ITCAM for Webspher is used around this type of problem E to quickly identify, locate and solve problems in applications. It needs to be stated that the main issues discussed here are those related to the application, while the performance tuning of the middleware, such as Websphere itself, is not covered in this article.

Description: ITCAM for WebSphere is an IBM-provided management and analysis software designed to manage, analyze, and diagnose composite application issues, which can help users quickly locate performance bottlenecks and identify errors and problems in the code and server resources that cause application problems, ensuring that applications run smoothly.

Methods of applying problem analysis and diagnosis

Before introducing the application analysis using ITCAM for WebSphere, let's take a look at the methods of applying problem analysis and diagnostics and compare them.

There are several ways to apply problem analysis and diagnosis, depending on the emphasis, there are usually the following methods:

Compare Java objects

This approach is typically used to analyze files such as Java core after a problem with Java EE applications. A comparison of the Java object information from the Core Dump to see which objects have abnormal growth and resource consumption at specific times. In this way, we can find the problems in the program development, or find the bottleneck points during the operation to optimize.

Benefits: WebSphere Technical Experts are more commonly used for this analysis.

Disadvantage: This method of analysis of the process is more complex, requires a higher professional knowledge, and this approach is mainly suitable for ex post analysis.

Analyzes the pending requests that are running and the Java stack objects and methods associated with this request

This approach is generally effective at a time when Java EE application problems occur, capturing the requests that occurred at that time, and the invocation hierarchy of this request to trace to a method of a class.

Advantages: Analysis of online application problems are most effective.

Disadvantages: There are fewer tools available for online analysis, and the analysis tool itself has the resource consumption for the middleware.

Set up alarms to diagnose the root cause of the problem by analyzing the invocation relationships and Java stack objects and methods that are invoked when the alarm is generated (such as slow or abnormal applications). Usually setting alarms is only a warning, so in order to effectively analyze the situation when the problem occurs, this method requires the collection of detailed application site information at the same time as the alarm, including from a request to enter a layer of information drill, until the lowest method call