When it comes to anything related to information technology, it’s always the little things that ultimately kill you. And nothing can kill a managed service business faster than all the costs associated with tracking down some intermittent performance issue that cannot be easily replicated.

Anybody running a managed service, or for that matter anybody in IT, has seen their share of gremlins, otherwise known as those inexplicable issues that adversely affect the performance of an application. There’s always a rational explanation. But how much time it will take to ultimately determine the cause and then fix it is anybody’s guess.

Michael Corey, the CEO of a company called Ntirety, a provider of remote database administration services based in Dedham, Mass., dealt with these types of issues when his company first ventured into the managed services space.

It didn’t take long for Corey and his colleagues to figure out that they needed a structured approach to dealing with application and database performance issues. Otherwise, they would have spent all their time chasing gremlins at the expense of making a profit.

To deal with this issue, Ntirety turned to a vendor called Precise Software Solutions, which started out life as an independent company before being acquired by Veritas, which in turn was acquired by Symantec. Now Precise has been spun out of Symantec as an independent company again.

As managed service providers of all stripes look to move up the proverbial stack in search of higher margin products and technologies, they are increasingly going to encounter random performance gremlins. Each one of those gremlins won’t kill the business all by itself. But collectively, gremlins can exact a heavy toll on your IT service department, which in turns adds costs and saps productivity.

Worse yet, new technologies such as virtualization and latency-sensitive applications built using Web services technologies will add overhead and complexity to any service that is dependent on managing applications. That makes all the more critical for solution providers to have a structured approach to dealing with application performance issues.

After all, each customer has their own unique issues. But the odds are more than good that if a customer has one problem, chances are another one will encounter a similar issue. It doesn’t make sense to constantly reinvent the wheel for each customer when the intellectual capital needed to solve the problem could already be residing in an overall management framework.

When it comes to managed services, the name of the game is to provide the greatest amount of automated service at the least cost possible. But providing excellent customer service can also drive you to distraction because of the costs involved. So at the end of the day, the only way to having a fighting chance against being driven mad by thousands of gremlins eating away at the performance metrics is to create a system to monitor application performance that can quickly and easily identify and isolate issues across all the elements that make up the application. Anything short of that is going wind up being little more than a one way ticket to the IT funny farm.