Working with HP Diagnostics
Author: Chuck Ezell | 3 min read | April 16, 2014
We’ve introduced many different ways in which you can profile your application code. There’s a particular tool Datavail’s Chuck Ezell has used to proactively profile code: HP Diagnostics.
The tool allows users to be proactive and returns very useful information in an easily digested visual format.
With LoadRunner’s load testing capabilities, HP Diagnostics allows a user to undertake both a full application analysis and a server diagnostics analysis. The user can also use Java profilers or probes that can be configured to save the results for eventual analysis.
Being Selective a Must
Ezell says because this tool can consume too much of a server’s resources, a user needs to carefully select the environments in which they should be used.
The instrumentation process also poses another particular challenge: If specific problems you’re monitoring in the profiling process don’t appear, you could think the code is fine when it isn’t, says Ezell.
The issue could be that it may not have been correctly instrumented. Instrumentation poses several challenges to users, claims Ezell.
Starting up Can Be Difficult
The initial process of using HP Diagnostics can be daunting, if not sometimes confusing. Multiple windows pop open during the process, requiring a user to determine which are being used to probe and to monitor. The more probes used, the more windows and more complicated the process can be.
Each Java probe used can take up resources. The more probes employed, the greater the CPU consumption and the more computationally intensive the process can become.
Drilling Down to the Root Cause
There often are challenges associated with drilling down to the root cause of the problem or problems. The database administrator could see that there has been JVM heap growth as a result of garbage collection, for example. What is causing this? Is it a specific user? Which transaction is involved in the process? These are the sorts of questions a user must understand, says Ezell, to be able to find the root causes of code problems. Sometimes it can be difficult to determine the application context.
But this sort of tool is invaluable, Ezell says:
This, he says, allows a user to answer questions and immediately identify real problems and patterns if you’ve instrumented it well.
And isn’t that what proactively profiling your application code is all about?
Do you need help in profiling your application code or better maintaining your organization’s database? Datavail can help. Contact us to find out how our experts can customize a solution for your organization’s unique challenges.
Image by bowie15/123RF.
