How to Deliver High-Performing, Business-Critical Web Applications - Page 2

Habit No. 1: Set and measure SLAs on business processes

Effective business processes are critical to achieving consistently superior services aligned with business goals. Accordingly, it's paramount to identify the business processes that matter most to your business, set and measure SLAs on those processes, monitor them 24/7 to consistently evaluate transaction success, and report results on a regular basis.

In doing so, you can understand the business impact of application performance, improve your ability to present relevant performance information in terms of the business, and bolstering the level of engagement between IT and business stakeholders. And, by providing a consistent approach to measuring business process SLAs where key criteria and performance metrics can be understood from a variety of perspectives, a consistent methodology and language between business and IT management can develop.
Habit No. 2: Monitor 100 percent of all user transactions, all of the time

To gain an accurate and complete understanding of application performance, it's important to manage 100 percent of all business transactions end-to-end 24/7-from the browser to the backend-as they traverse the complex, multi-tiered infrastructure. By monitoring real user transactions in production environments, you obtain valuable insight into the user experience and transaction success/failure. This allows you to quickly identify, triage, prioritize and resolve problems before your customers and business are affected.

You can manage all transactions in real time and map them to specific infrastructure components to pinpoint the source of problems quickly, eliminate finger pointing between IT operations teams and significantly reduce Mean Time to Repair (MTTR). In addition, managing application performance from the user perspective makes it possible to assign priority levels to users and organize users into groups to measure performance based on business priority.
When evaluating an APM solution, look for vendors that deliver a complete, fully integrated solution that combines real-time user experience management with business transaction monitoring in complex production environments-and delivers this with little or no impact to the IT environment. While synthetic transaction monitors offer complementary functionality, they are limited in that they do not reveal what real users are experiencing in production. As a result, they cannot tell you how your system is executing against SLAs or business goals, and cannot map the user experience to the infrastructure components.
Habit No. 3: Employ predictive and proactive monitoring

Most enterprise-class IT infrastructures are highly complex, heterogeneous and distributed, which presents unique challenges for monitoring business transactions. In these environments, even small incidents-such as depletion of threads and resources, memory leaks, changes or errors-can have a significant impact on overall application performance.

To manage this, baseline monitoring and heuristic-based trending technologies provide an added layer of predictive and proactive analysis that enables you to identify and alert on problems before they impact users. These technologies also identify internal failures and the application's interaction with the back-end systems to identify in-flight problems before they result in serious performance and availability issues.
An effective APM solution monitors all of this and more. It can even automatically notify leading system management solutions to integrate application performance insight into the broader infrastructure management environment. This can result in faster problem resolution, improved troubleshooting and streamlined management.


Habit No. 4: Prioritize incidents based on business impact

Once problems are identified, you need a way to prioritize them based on their importance to the business. By assigning business value to successful and unsuccessful transactions, you can prioritize incidents based on the importance of the user, criticality of the transaction and the severity of the issue. This provides you the real data that you need and a foundation to resolve the most business-critical issues first.