The promise of a monitoring solution that will pinpoint application problems and give you exact steps to fix the problem has remained a dream. In addition, monitoring systems have become notorious for being expensive and difficult to maintain. Diagnosing application performance problems requires application-specific diagnostic information that general-purpose monitoring tools often do not provide.
While system monitoring products are useful in triaging a problem and assigning responsibility to a particular team (for ex: Application Server team), they often do not provide the necessary details to help you determine the problem and fix it. Monitoring products are described by their users as mile wide, inch deep – great for providing high-level visibility into broader systems such as browsers, web servers, app servers, network devices, databases, storage etc, but not so great for specific diagnostic information that you need for fixing problems.
Instead, it often takes specific diagnostics tools tied to the application container to really be able to effectively drill down into the data sufficiently.
In this article, we will use Apache Tomcat as an example, and explore a few scenarios where Tomcat administrators need more information to help determine the problem.