New Relic has unveiled a suite of AI-powered services under the brand New Relic Applied Intelligence (NRAI), aimed at helping enterprise customers quickly discover and resolve crucial operations issues.
NRAI analyses web application data to automatically identify any abnormalities and then provides engineering and operations recommendations on how to resolve them before they impact customers. The intelligent tool can also predict potential problems and prevent them from developing into major issues.
"We're in the business of developing applied intelligence, and we're baking it into our entire platform," New Relic CEO and founder Lew Cirne announced at the company's FutureStack conference in New York. "This isn't one feature, this is going to be everywhere throughout everything New Relic does."
This is similar to the range of machine learning capabilities New Relic rival Splunk rolled out in September last year. Read next: Splunk brings machine learning capabilities into its tools and launches toolkit for customer’s own algorithms
NRAI is built into New Relic's SaaS application performance monitoring (APM) platform and driven by the trillions of event data points that New Relic processes from its customers' critical systems every day, all underpinned by the cloud.
"We have our customers continually come to us asking for help to make more sense of the data," says Cirne. They never come to us and say, New Relic, I wish you would build an AI."
Instead, they ask the company to deliver on three specific needs: to filter through their data to highlight the most important aspects, to prevent problems from occurring, and if those problems can't be stopped, to fix them as fast as possible and with the minimum disruption.
What is New Relic Radar?
New Relic launched three new NRAI features at FutureStack: Radar, NRQL Baseline Alerting, and New Relic APM Error Profiles.
Radar leverages artificial intelligence to deliver proactive recommendations to DevOps teams.
It provides a personalised feed of predictive and prescriptive insights into software performance. Data analysis allows it to automatically detect anomalies in performance patterns and make suggestions on how to fix them and what to prioritise. Each individual Radar user receives personalised recommendations based on their individual requirements.
These recommendations are presented to individuals on single cards, categorised as Advice, Perspective, Events and Celebrations. That staff member can then share the card with the rest of the operations team through Radar’s Slack integration, for example.
Radar started as a research project called Project Seymour within the company. New Relic ran it on their own systems last summer and previewed it at the previous edition of FutureStack, back in November 2016.
The benefits they observed in their own complex environment, and the feedback they received from customers at the demonstration, convinced them that it was worth developing into a consumer product.
They rolled out a beta version of Radar in December and are now releasing it to paying customers.
Other NRAI services
Another aspect of NRAI is Baseline Reporting. This sets thresholds based on historical transaction and event data that generates customised alerts.
The company has extended this to New Relic Query Language (NRQL) Baseline Reporting, which lets users receive alerts based on any query written in NRQL.
"Our competitor products baseline like once an hour," says Cirne. "We're baselining in minutes, so that you can see super highly-sensitive metrics and then you can customise what you want to baseline, as easily as you write an NRQL query."
Error Profiles completes the trio of new NRAI features. It analyses error attributes in comparison to historical values and, with a click of a button, gives the user a view of everything unusual or unique to that error.
New Relic has also developed an inter-connectivity solution that provides visibility into complex environments of multiple services, such as the complex web of interacting technologies that allows a tourist to easily find and book a room on Airbnb, one of New Relic's more than 15,000 customers.
The feature is known as Distributed Tracing. It can track user request through every service in the chain to make it easier to discover any performance issues. The tool is currently in beta, but Cirne said he wants to ship it "as rapidly as possible".
Go fast at scale
These intelligent insights are part of New Relic's objective to help companies "go fast at scale".
"Back in the day we built software in a very traditional way," says Cirne. "We shipped releases. Monolithic releases."
"Shipping" software in those days had a literal naming, as annual updates were burned onto CDs and sent to customers by sea.
The software of today is delivered in a very different way. Cirne compares the method to an ever-flowing stream of deploys, each of which adds extra drops of value.
As companies like Airbnb ship updates around 180,000 times per year, this increasing speed brings an element of risk. New Relic mitigates this with instrumentation, automatically monitoring applications while they run in production so users can rapidly react to and rectify problems.
"If it's not instrumented, it's not production in a professional sense, because how can you possibly move fast if you can't see the thing?" asks Cirne.
According to Gartner, 95 percent of applications running in production today are not instrumented. Cirne believes this is recklessly risking the health of the applications. He wants every aspect of the software lifecycle to be easily and cost-effectively instrumented, from entering production opening up to for business and keeping it running forever after, stably and securely.
"It's possible to go fast by strapping a rocket onto your go-kart, but that's not the goal," says Cirne. "It's not about just moving fast. It’s about moving fast with confidence. That's the goal."