New Relic

It is 2:14 a.m. somewhere, and a checkout page just stopped checking out. Revenue is leaking by the second. Before the panic spreads, an engineer opens a single screen, follows a slow request from the browser down through a microservice into a struggling database, and finds the culprit. No war room. No guessing. That screen, more often than not, belongs to New Relic - the company whose entire reason for existing is to make sure nobody has to find out their software is broken from an angry customer first.

New Relic sells visibility. Not features, not dashboards for their own sake - visibility into the messy, distributed, half-understood systems that now run banks, hospitals, retailers, and games. The industry calls it observability. New Relic helped name it.

Software got complicated. Nobody could see inside.

For most of computing history, software lived on a server you could walk up to and kick. Then it dissolved. It moved to the cloud, split into hundreds of services, scaled up and down on its own, and started talking to other software nobody fully controlled. The result was a strange new problem: companies were betting their business on systems that no single person could actually see.

When something broke - and it always breaks - the response was archaeology. Engineers dug through log files, pinged colleagues, and rebooted things hopefully. Downtime was measured in hours and apologies. The tension New Relic exists to resolve is simple to state and brutal to live with: the more important software becomes, the harder it is to understand.

A second act, and a clever anagram

In 2008, Lew Cirne made a bet that engineers would pay to see inside their own software - if you made it easy enough. He had reason to believe it. Cirne had already built Wily Technology, an application monitoring company acquired by CA Technologies. New Relic was, in a sense, his rematch with the same problem, now armed with the cloud and a software-as-a-service model.

He also had a sense of humor about it. "New Relic" is an anagram of "Lew Cirne." It is the rare enterprise software company named by shuffling the founder's own letters - a small, slightly Wildean joke hiding in plain sight on millions of invoices.

FILE PHOTO: A company name that is literally its founder, rearranged. Branding consultants hate this one trick.

The bet paid off. New Relic made instrumentation almost embarrassingly simple - drop in an agent, and within minutes you could watch your application's performance in a browser. Engineers adopted it from the bottom up, often before their managers had heard of it. That motion - developers first, purchase orders later - became the company's growth engine.

One screen for the whole stack

The pitch is unification. Most teams used to stitch together a different tool for each layer - one for the app, one for the servers, one for logs, one for the website. New Relic's argument is that the layers are connected, so the tools should be too. Everything lands in a single telemetry data platform, queryable through a SQL-like language called NRQL, which turns a wall of charts into something you can actually ask questions of.

There is also the matter of the bill, which is where New Relic made a genuinely contrarian move. Instead of charging per server - a model that punishes you for growing - it shifted to usage-based pricing: pay for the data you send and the people who log in. A free tier lets a curious engineer start for nothing. It is a pricing page that, refreshingly, does not require a sales call to understand. The wager is that if the tool is good and the entry cheap, adoption spreads on its own merits rather than on a salesperson's persistence.

None of this works without context, which is the quiet word doing the heavy lifting. A slow database query means little on its own; a slow query tied to the exact line of code that called it, the customer who felt it, and the deploy that introduced it means everything. New Relic's bet is that connecting those dots automatically - rather than leaving humans to do it under pressure - is the difference between a tool and a teammate.

The numbers behind the argument

Skeptics are right to ask whether any of this matters at scale. The receipts are reasonable. New Relic serves thousands of companies across finance, retail, healthcare, media, and government, and as a public company it reported around a billion dollars in annual revenue. Its 2023 take-private deal valued it at $6.5 billion - shareholders received $87 a share, and founder Lew Cirne rolled a large slice of his stake back into the private company rather than cashing fully out.

It earns analyst nods, too. In 2025 New Relic was named a Leader in the inaugural IDC MarketScape for Worldwide Observability Software, and a Leader in the Asia Pacific AIOps assessment. Its cloud partnerships span AWS, Microsoft Azure, and Google Cloud, and it leans into OpenTelemetry, the open standard for instrumentation - a notable stance for a company that could just as easily lock customers into its own agents.

Debug with data, not opinions

Strip away the product names and New Relic's mission is almost stubbornly plain: help every engineer create more perfect software. The phrase "using data, not opinions" shows up everywhere, and it is more than a slogan - it is a worldview. In a world where the loudest voice in the incident channel often wins, New Relic's pitch is that the data should win instead.

The agents are coming, and they need watching

Here is the twist that keeps the company relevant. The same AI that promises to write and run our software also makes that software harder to understand. Agents call models, models call tools, tools call other agents, and the whole chain becomes a black box even its creators cannot fully explain. New Relic's response is to point its telescope at the AI itself - tracing LLM responses, comparing models, and in 2025 opening a partner ecosystem built specifically for agentic AI. Its own AI-monitoring usage has been climbing roughly 30% quarter over quarter, which is either a trend or a tell.

So return to that 2:14 a.m. checkout page. The version of that night without observability is the one every engineer remembers and dreads: hours of digging, a furious executive, customers gone by morning. The version with it is almost boring - a query, a root cause, a fix, back to bed. New Relic's whole business is the unglamorous work of making 3 a.m. boring. In software, boring is the highest compliment you can pay.

Interviews & demos