The current wild pace of enterprise technology development has a host of enablers: devops, cloud platforms, microservices, rampant open source code-sharing, and so forth. Application monitoring sometimes gets left off that list, but it shouldn’t be -- you need a deep view of how applications perform in production to determine how to improve them continually.
“Full-stack visibility” is how New Relic describes its ability to give operations, developers, and even business execs a deep view of how applications are performing in real time. During an InfoWorld interview last week, to illustrate the volume of monitoring data collected, CEO and founder Lew Cirne fired up a New Relic smartphone app: “At this moment we’re collecting 3.1 million events per second ... and in the last minute we queried 5 billion events per second,” he said.
That little demo neatly demonstrates the ability of users to access application performance data instantly from New Relic’s cloud-based service. Cirne deserves credit for convincing investors and customers that a fairly geeky IT function, APM (application performance monitoring), should be a first-class application for taking the temperature of digital business.
Along with offering the first SaaS APM play, New Relic’s secret sauce has been a compelling user experience, with configurable dashboards that present performance data in real time. That approach kicked up a notch last November when New Relic added business intelligence style features to its Software Analytics Cloud.
Riding the digital transformation wave of the past few years, Cirne saw his company achieve unicorn status shortly before its IPO in late 2014. Last week’s earnings report showed $52.5 million in revenue for the quarter, up 57 percent year over year, on an installed base of more than 1,500 paid enterprise business accounts.
New Relic uses agents to monitor server-side applications in the cloud and on premises, but its tendrils extend all the way to mobile and browser-based apps. The company made its initial splash as a solution for cloud-based applications, yet Cirne claims that customers are increasingly using New Relic to monitor on-premises applications as well. One thing for sure, says Cirne, is that legacy APM can’t make the leap to monitor cloud applications effectively:
One of our customers in a very large, multi-thousand [cloud] server environment has 20 percent of their servers come and go any given day. If you have old fashioned tools that alert you when a server goes down, goes away, that’s not very useful -- it’s awfully noisy. ... That’s even further complicated when you have Docker and containers. Some of them we call “mayflies.” They’re in existence for minutes, and they go away.
But the biggest difference from old-school APM is in who’s looking at the data. Last year, says Cirne, Gartner did a study which showed that the average APM implementation cost $800,000 and had 3.5 users. Widening that audience is central to New Relic’s play:
You get the digital business leaders, you have all the developers, all the operations people all on the same page with the same data. Their view of the data may be a little different: Developers may be looking at stack traces, ops people might be looking at aggregate service levels, and business people might be looking at how many widgets I sold in the last hour ... We have customers definitely showing with our software the correlation between adding 300ms to the page load time and the business impact, how many fewer items are sold when that happens, and then being able to diagnose it.
Delivering performance data to developers in particular can have dramatic impact, especially in modern devops and microservices environments that have so much more to monitor. Continuous delivery and deployment of granular services increases the risk that any one service may fail or perform poorly and affect multiple applications. Along with monitoring all those microservices, New Relic can map the dependencies among thousands of them, says Cirne.
The company has also been quick to jump on the Docker container bandwagon. “We believe the container is a first-class citizen in that it has to be part and parcel of the application monitoring solution,” says Cirne. Interestingly, he says he wouldn’t rule out getting into the container management space. With performance measurement at the container level, playing a role in optimizing workload placement across a wide pool of servers seems like an intriguing next step.
The past several years saw a surge in startup funding -- what Keith Rabois of Khosla Ventures has called the “steroid era” of startups -- and New Relic was clearly a beneficiary. A sharp drop in the company’s stock price earlier this year may indicate that expectations for APM and its associated analytics may have been a little oversold.
On the other hand, our most recent tech boom had an underlying message that stuck: Internet applications have become an absolutely vital part of business. To improve them on an ongoing basis, you need not only monitoring, but also the ability to analyze and present meaningful information to multiple stakeholders. New Relic will see increasing competition from AppDynamics, Dynatrace, and others in this space, but the desire of enterprises to ingest and analyze gobs of app data is only going to increase.