June 2, 2015

Monitoring the Performance of Widely Distributed Applications

Written by Gen

keyword icon Uptime Monitoring

More and more system components exist outside the IT infrastructure, and more users are accessing your company's apps on mobile devices and in other unpredictable scenarios. Today's monitoring systems are keyed to the business users who create, collect, and consume the data. You need a monitoring service that accommodates the many different ways your apps are being used by the people inside and outside your organization.

Today's systems integrate components from an ever-growing range of sources. For example, the typical web app combines dynamic data from several internal and third-party sources: from databases hosted in house and online, to partner content, to widgets from online ad networks and social media.

In an April 29, 2015, article on Forbes, Despreneur.com Editor-in-Chief Tomas Laurinavicius cites figures from Etsy that found the 40 percent of visitors abandon a site when it takes longer than three seconds for the page to load. But to diagnose the cause of slow page loads, you need to measure the many components that comprise a single transaction.

For example, slow database-reopen times are often the source of poor page response. That's why you need a monitoring service such as Happy Apps that can measure the performance of your database servers and your web servers, among other relevant resources.

Load-time tests conducted on 2,000 e-commerce sites by Strangeloop Networks in 2014 indicate little patience among site visitors. Compared to business sites that load in 1 second, pages that take 3 seconds or longer to load experience a 50 percent higher bounce rate, 22 percent fewer page views, and 22 percent fewer conversions. The test results are reported by Aykira's Keith Marlow in an April 4, 2014, article.

If forced to wait three seconds or longer for a page to load, 57 percent of visitors will abandon the site, and 80 percent will never return. Source: Strangeloop Networks, via Just Creative

Monitoring the components that comprise the stack

Initially, cloud-based system-monitoring products were platform- and server-specific. Nari Kannan explains in a TechTarget article that monitoring agents were separate from the apps, databases, and other systems they were reporting on. Now agents are integrated with the systems they monitor, and you can assign them to specific instances. To represent system performance from a user's perspective, you have to monitor the various distributed components: the public and private network, the user interface, and the app or database itself, among others.

Erwan Paccard writes in an April 22, 2015, article on SD Times that the only way to represent all the many pieces of the distributed-application puzzle is by working with the people who create, collect, and analyze the data. These "digital business owners" provide insights about customer expectations, which is the true driver of today's data networks. In particular, the shift to mobile platforms sharpens the need to deliver fast, effective interfaces that don't bog down the user experience.

Perhaps the most difficult variable to measure is the performance of the public Internet, particularly when serving data from ad networks and social media. In an April 24, 2015, article on Business World, Bill Ting cites a Gartner survey that found 70 percent of performance problems are reported by end users.

Monitoring web applications requires collecting and analyzing data from application and database servers, load balancers, and other sources. Many of the instances that comprise the web-app stack reside in the cloud, outside the traditional IT infrastructure. Tying it all together for users is a crisp, clean dashboard that delivers a complete-yet-comprehensible view of the hybrid network.

The basic web-application stack is comprised of application servers, database servers, load balancers, and other instances, all of which must be monitored separately. Source: Amazon Web Services Documentation

All the elements of a distributed-monitoring solution come together in the Happy Apps service. Happy Apps is a robust app-management service that supports SSH and agent-based connectivity to all your apps on public, private, and hybrid clouds.

All checks performed on your systems are collected in easy-to-read reports that can be analyzed to identify repeating patterns and performance glitches over time. If you're looking for ways to save time, trouble, and money when managing your apps, databases, and other systems, visit the Happy Apps site for pricing information and to sign up for a free trial.

comment icon Comments

* Required