DX Application Performance Management

APM Blog: Problem Solving: Six Ways to Avoid the Wrong Path

By Hallett German posted 08-16-2015 01:08 PM




In Colonial Williamsburg near where I live, there is a maze in the garden behind the Governor's Palace. At the start of the maze, the sign invites you to go left. Once inside, you will get easily lost and misdirected. But eventually you will find your way out.

At the end you discover, if you went right in the first place, you would already be by the exit.



Sometimes, I see similar behavior in APM problem solving where if only someone asked just one more question or did just one thing differently, then the issue would be resolved already.



Six Ways not to get Trapped.



Here are six ways to avoid being misled from least to most importance.


1. Ignoring False Positives

Sometimes log messages look like warnings or errors but are really information messages. This is especially true when turning on DEBUG or VERBOSE mode in Introscope logs.

Knowing if something is really an error worth pursuing or not can help maintain proper focus.


2. Removing Incompatible software

Often, an issue is found due to software  incompatibilities. Such as an EM release being later than an agent release. Or a JRE release that is unsupported in the current APM release but is supported in a later one. Checking the compatibility guide is a quick task to do. Please do this early not later with a problem as to not to delay resolution.

3. Find out when the issue started.

Often, customers have a good idea when a behavior or error began to manifest. However, what they might not be aware of is the issue started long before that. So reviewing logs may get to the real state date. Once there, the behavior may be tied easily to some event.

4. Discover if something is a symptom or a problem.

Many issues may be a symptom of something else such as performance issues, reaching server capacity, unoptimized software and more. Peeling the layers of the onion to get to the real issue may take time. But in the long run will bring about system stability and scalability.


5. Not checking if this is a Known Issue

A five minute check of the release notes for your current and future releases, as well as the always updated knowledgebase may quickly indicate if something is a known issue or not. This will help avoid hours of aggravation.

6. Determine if there are Cascading Issues

Recently, there was a customer had an issue where data was not leaving the TIM. So, the focus was placed there. However, the real culprit was the Tim Collector. It  was not doing its job since it was performing other services (A common issue). As noted in the Tech Tip below, knowing the functional workflow of a task including likely failure points and what is next in a successful workflow can be very helpful.



I hope this is helpful and aids you in avoiding future issues. Please respond to this blog as to other things that can misdirect a problem resolution effort.