Performance means different things to different people when used in terms of software applications. It must have some context for better understanding. Application performance is measured against two sets of performance metrics. The actual performance observed or experienced by the application users remains one of the most important metrics for measuring application performance. It includes the average response time during peak and normal loads. Measurements related to average response time include the time taken by the application to respond to a user's action, such as a page refresh, navigation, or a button click. They also include the time taken to perform certain operations, such as sorting, searching for, or loading data.

This section is meant to provide technical teams a perspective on some of the aspects of configurations and internals that can be set or altered to optimize effects, in order to improve the performance of the application. In usual cases, technical teams never keep an eye on the memory that the application uses or the CPU utilization unless they are stuck with a performance issue. Application transactions include requests received by the application per second, database transactions per second, and pages served per second. The load on the system is usually measured in terms of volume of transactions that the application processes.

There is another set of measurements that involves measuring the computational resources utilized by the application while performing operations. It is a very good way of identifying whether the application has enough resources to sustain the given load. It also helps in identifying whether the application utilizes more resources than it is expected to. If so, it can be concluded that the application is not optimized on the performance side. Cloud-hosted applications are popular these days. In this era, it is important for users to have the same experiences on applications deployed over the cloud or a non-cloud infrastructure, and on the local environment.

Application performance monitoring and improvements may not be necessary for an application, as long as it performs per expectations. However, as part of the application development life cycle, new requirements come up, new features are added, and the application becomes more complex by the day. This starts impacting the application's performance, as the main focus is kept on new feature development. A time will come when the performance is not up to marks, because no one actually works on application performance enhancement.