On Demand Replay Now Available
All The Leaves Aren't Brown: Many Ways to Profile Your Application Code
Finding improvements in your Client Side JavaScript, Java or .NET code is often easier said than done. Managing application performance problems due to memory leaks, poor coding habits, superfluous library usage, detecting deadlocks, measuring memory usage and garbage collection frequency are activities often left to the DBA and System Admin. Once the code has left the developer's desk and hastily run through some degree of QA, it's often pushed to production. Often, the DBA doesn’t know the code or even the developer, and fixes must be derived in the dark. These tasks don't have to be stressful and scary if you have the right tools and know how to implement them. These tasks don't have to be stressful and scary if you have the right tools and know how to implement them.
Profiling application code goes far beyond measuring response time and working out the bugs. In addition there are different methods and types of profiling application code. Even greater still, the degree of code being deployed in today's applications and how it varies in implementation is enough to cause anyone to throw their hands up. Do we concern ourselves with the stability of the JVM on a user's computer or just the application server? Does it really make a difference what web browser the user chose to use that day? How do we know how the difference between a locked thread and a locked database object even though they seem to look the same from the users perspective? We'll discuss issues like these digging into the application and client tiers to find some sanity.