When it comes to staying on top of social and digital media, web culture, and technology at large, there’s no better or more prolific source than Mashable. Founded in 2005, it’s the premier destination for breaking web news, analysis of technology trends, reviews of new websites and services, social media guides, and more. Mashable gets an astonishing 40 million monthly page views from a diverse, information-hungry audience — early adopters, social media enthusiasts, entrepreneurs, influencers, brands and corporations, marketing, PR and ad agencies, Web 2.0 aficionados, technology journalists, bloggers, tweeters, and Facebook users. The publication deploys PHP and Ruby on Rails apps, both public and private, using an open source back end and stack. When it comes to app management, the primary objective is always ensuring an optimal user experience.
Brian Doll checks in with Mashable CTO Frederick Townes to discuss web performance:
News is all about timing, so being in the lead with a story and living up to the end-user expectations of up-to-the-minute reporting is critical for Mashable. The engineering team, therefore, must be free to pay close attention to the UI. As CTO Frederick Townes explains, “Our goal is to maintain the highest possible Apdex scores so we can confidently focus our efforts on the user experience of our applications, rather than on scale or performance anomalies.” (Apdex is an open standard that defines a method to report, benchmark, and track application performance from the end-user perspective). With the business moving at lightning speed, Townes and his team didn’t have time to develop their own APM solution while supporting and extending current administration tools and attending to every other pressing issue. So, while making changes to their stack, they realized the transition was a perfect opportunity to continue improving the application environment. “We needed to be able to easily identify – at a moment’s notice – what’s going wrong in an app before our internal or external users reported it,” recalls Townes. “We were also looking to gain more control and better alert capabilities with a flexible solution that made it simple to segment parts of a single app within the instrumentation.”
Staying Ahead of the Action
Prior to the transition, Mashable’s engineers used a variety of open source tools at different points in the stack for regression tests, unit tests, and profiling. They didn’t replace these utilities, because each served a unique purpose. But they decided it was time to implement a complete, on-demand APM solution for their production environment. And that’s when they turned to New Relic RPM. After a quick, painless deployment, RPM drastically reduced the time required to monitor their complex stack and multiple applications. Townes reports, “In less than an hour New Relic was pushing data into the instrumentation for all of our apps. Then what really made RPM an asset for us and, by extension, our users, were the comprehensive functionality and elegant integration with our stack and all our tools. Even with the large number of machines we have, RPM makes it far easier to track down issues caused by edge cases (which any developer will tell you are the hardest to pinpoint) because it provides early warnings and helps us see what’s happening with the apps in great detail.” In fact, the solution’s alerts and level of error detail are two of the features he finds most valuable and spends the most time using.
In a recent case, for example, errors were popping up on one machine and not another. RPM helped Mashable identify the root cause right away, whereas the old debugging methods would have cost an excessive amount of time, which could have spelled trouble. Moreover, the solution allows Townes and his team to make changes to individual applications or the stack, then ensure everything is equal across the board by using the same policies for each machine to monitor performance and check for errors. “We save so much time on administration and maintenance now, which really allows us to focus on the user experience,” says Townes. “The big key for us is greater insight and awareness. It’s very important to be able to dig deeply to understand the apps’ behaviors and patterns, either through raw data or visualizations. Doing that manually works, but it’s far more time consuming and generally less scientific. RPM gives us the control to drill down into the detailed data and still get a comprehensive view of overall performance.”
Mashable Gets the Scoop on Expansion Without Stress
For the engineers at Mashable, the ultimate benefits of the enhanced capabilities afforded by their New Relic solution are peace of mind and a stronger product. For the company’s ever-growing audience meanwhile, the engineers’ ability to rely on consistent, seamless application performance and focus on user needs ensures an excellent experience every time they visit. “Having RPM in our arsenal now makes it much less stressful on us to run a complex stack with lots of apps and mitigate all the things that could go wrong,” says Townes. “It also helps protect our users against performance issues, which keeps them happy and makes life better for my team as well.” The site and its audience are sure to keep growing rapidly, and he anticipates RPM will continue to aid the expansion as they move forward. “It’s a strong product, very easy to use and also well supported. There are a few specific features we plan to start taking advantage of soon, like the customization options for instrumentation and reporting. In general, I definitely see it as an increasingly invaluable tool for us in future efforts.”
Be in the Know
Mashable is everywhere, giving you access to the latest news in social and digital media, technology and web culture however you prefer to connect: Facebook, Twitter, YouTube, StumpleUpon, Digg, and via apps for Android, iPad and iPhone. You can also visit the flagship site at http://mashable.com to catch up on the day’s top stories, submit news and tips, or find out how to advertise with them.