Don't use a crystal  ball and guess the cause of application performance problemsI used to have a co-worker who had sayings for all occasions, but one of his favorites was “When the birds sing, the flowers bloom”.  The point he was making is that you need to be careful to not make a cause and effect correlation between two events, even though they seemed to be related.  In his example, the point was that it wasn’t the birds singing that caused the flowers to bloom.  It was another common attribute, spring, which brought the singing birds as well as caused the flowers to bloom.

When there is a problem, you can’t try to guess what caused the problem and where the root cause is, but we see people do it all of the time.  For example, you’ve probably heard things like:

  • “The payment application is slow and we see that the JVM memory is at 90%. That must be the problem.”
  • “The purchasing application is not responding.  When we had this problem last year, the problem was in DB2.  Can someone reach them to see what the problem is?”
  • “We’ve seen a spike in the messages going to the Dead Letter Queue.  They rolled out a new version of customer lookup last night.  We need to roll back the changes.”

Unfortunately, while the intent behind all of these ideas is well placed, they are most likely examples of coincidence.  You could have just as easily blamed the birds for all of these problems (“Payroll processing is slow… and the birds are singing. That’s the problem.”)

The key to solving complex application problems lies in using a technology designed for solving complex application problems.  You need to leverage a tool that can correlate data from multiple sources, one that can identify trends, and can quickly determine the root cause.

That is the reasoning behind AutoPilot® and its underlying complex event processing engine.

To find out more information on how to get better visibility and early warnings about application performance monitoring issues, please check out this free whitepaper, Unraveling the Mystery: How to Predict Application Performance Problems