New Relic aims to be your dashboard of the future

Promises to keep private and public and cloud software in synch

new relic ceo lew cirne2
New Relic

In Lew Cirne's view, all companies are now software companies and understanding how your software is treating your customers is key to business success. Cirne is the founder and CEO of New Relic, a cloud-based provider of application management tools. In this CEO Interview Series conversation with IDG Chief Content Officer John Gallant, Cirne explained how New Relic gets IT and business execs on the same page in improving operations and customer experience, and he described the company's new tools for keeping highly virtualized private and public infrastructure in synch. He also talked about a 'unique' pricing scheme that recognizes the dynamic nature of computing today, and outlined why existing management tool vendors have a long way to go to catch up with New Relic.

What hard problems do you solve for customers in this cloud world?

The primary thing driving cloud is that businesses are becoming software businesses. Software is becoming the lifeblood of companies. One of our larger customers is GE, which is well known and advertising itself as becoming a software business. We also feel that digital is different from traditional software. It is a team sport that involves not only developers and traditional IT but also DevOps and, increasingly, business people. They all need a first, best place to look to understand everything about their software: How customers are using that software, what is the customer experience and what is the customer activity in that software that drives business outcomes? If all of these stakeholders are on the same page, using the same platform to watch their software in real time, that turns into digital success and business success. At New Relic, our mission is to be the first, best place to look to understand your digital business and we do that for 14,000 customers today.

New Relic uses the phrase 'full stack visibility.' I want to understand exactly what that means for customers.

The application stack has changed dramatically, particularly in the last five years, as customers have moved from monolithic applications running predominately in Java or .NET to multiple programming languages running in containers, often in the cloud, and many microservices. The stack has changed and gotten a lot more complex and dynamic. We see containers that have lifespans of minutes in production where, historically, servers would run for years and the same process would run for months. We have more dynamic and interconnected application environments. It's more important than ever to capture everything about what’s going on in the software and the related infrastructure in real time and then to be able to make sense of it in a way that allows people to quickly understand whether there are health problems in that software and, if so, how to diagnose them. Beyond that, helping them understand the customer experience.

Sticking with that full stack visibility, what can you look at that a traditional application performance management tool wouldn't be able to?

Traditional application performance management tools have been very focused on one or two programming languages and they don't typically address other languages like JavaScript, Node, Python or Ruby. We see increasing heterogeneity in languages. You also need to go deep into the browser and to see when a page load takes seven seconds. Often, that can be not only the software you provide on your server but all these other services on the web doing ad tracking or other things. They can be impacting your customer experience. [You must] be able to track the depth of visibility in the browser and then going deep into the mobile device, which is the new front door of the business. Recently, we announced a new approach to infrastructure monitoring particularly well-suited for cloud deployments where we can track in real time the state and configuration of every virtual host running in your environment. We can instantly tell when you've got a variance in configuration or a change on an instance and then correlate that with the application health. We believe that all belongs in the same platform and in the same view that we provide.

How is the SaaS-based approach you take to APM superior to on-premises APM?

There are a few things that make it fundamentally superior and, as context, I was the founder of Wily Technology [a mainframe application management vendor acquired by CA Technologies in 2006]. I skinned my knees building a great, on-premise APM product. I learned enough from that experience to know that there are fundamental advantages to being a pure-play cloud company.

One is we are at massive internet scale -- a true multitenant cloud solution that collects about four million events per second, every second. We handle that with ease. In context, that's more than 100 times the number of tweets that exist in the world.

And then we are able to query. We've seen peak query loads of 12 to 20 billion events per second, in aggregate, scanning 20 billion historical application performance events per second from our cloud. That's just not feasible in any cost-effective way if you build that kind of horsepower on-premise. In the past, most APM products tried to catch up with the throughput of the application by pre-aggregating in anticipation of what questions might be asked, like what's the average response time of this application? Wily was good at answering what's the average response time of this application.

But if you break it down and you want to know the 98th percentile response time for this customer, to do that you need to capture every transaction that flows through every part of the service so you can just pick out the ones that a specific customer executed. It might be looking at a different customer or it might be segmenting in a totally different way, like signup date or feature or product category or EC2 instance types. We can arbitrarily segment in real time the health of the system because we don't aggregate the answer until the person runs the query. That is a far larger data problem than most of the applications that our customers can run in their data centers today.

1 2 3 4 Page 1
Page 1 of 4
It’s time to break the ChatGPT habit
Shop Tech Products at Amazon