Penguin Random House is using New Relic to monitor the performance of an array consumer-facing apps that have been developed to provide a digital adjunct to its vast catalogue of books.
"New Relic is critical," says Brian Uckert, who heads up continuous integration and deployment at the company and manages the monitoring of every application at the publishing house.
"It gives an insight to the developers as they're building their code and developing. It pinpoints bottlenecks and errors. Within their code, developers have access to New Relic and the dashboard and they can actually view slow queries specifically, which has been instrumental in keeping the development cycle agile."
The digital disruption to publishing
Penguin Random House is one of the biggest publishers in the world. It boasts more than 70 Nobel Prize laureates in its catalogue, and contemporary authors from Dan Brown to Richard Dawkins.
It publishes 15,000 print and 70,000 electronic titles annually, but like its competitors, has struggled to adapt to the digital disruption to their business. The rise of Amazon and eBooks has transformed publishing more than almost any other industry.
So a new range of consumer-facing apps are a way that they can capitalise on digital innovation, primarily by adding an interactive element to their books and authors. They can be powerful promotional tools and lead directly customer conversion through the inclusion of pre-ordering information.
Penguin Random House apps range from Game of Thrones companions with character profiles, maps, anti-spoiler protection and even Dothraki language courses, to Stephen Hawking-inspired interactive experiments exploring theories of the universe, and Dr. Seuss picture books with colouring tools and animated effects.
Uckert's team also develops apps to use in-house and gain a competitive business edge in areas such as prospective sales reporting on specific books and authors.
New Relic makes sure the apps are rolled out quickly and then run at their optimum performance.
They use New Relic to monitor every aspect of the app performance down to specific lines of code and the Insights analytics tool to present the raw data in real-time data visualisations about issues such as bottlenecks with the coding for International Standard Book Numbers (ISBNs).
"We have found Insights to be incredible. That has given our Java developers a way into their Java files to see specific problems and be able to create specific dashboards to them without pulling IT in for every piece along the way.
"We have some very strong developers and the developers can go in and quickly create a dashboard and have it living to see and follow up."
He also uses New Relic's new Applied Intelligence feature Radar to detect and resolve any issues which they might have otherwise not known. That integrates with Slack to provide convenient alerts for the developer team.
Why New Relic?
Uckert joined Penguin Random House around 18 months ago, and one of his first major decisions was to bring in New Relic to monitor app performance. He chose it for its easy deployment and fast results, and says its main competitor, AppDynamics, is not as agile.
“It's not a SaaS product," he explains. "It's very much a different product, so it probably has strengths but for what we need New Relic provides us a quick insight and that's really what we're after."
The intuitive design has led to high user adoption rates. Business managers at Penguin Random House also use New Relic to gain an easy understanding of app performance issues such as CPU usage by viewing the visual dashboards.
Earlier in the day of our meeting Uckert received a call from a project manager about a new release due in days that was experiencing slowness.
"I launched the dashboard and was able to pinpoint quite rapidly and then send a permalink directly to them so they could see exactly where the issue was," he explains. "It was quickly identified at 10.30 this morning as an API that was slow, and it was able to be identified and corrected."
"That took minutes. Probably about four or five clicks to identify, to send them a link and have them actually verify that as the problem and resolve it."
Engineers at the publishing giant have to roll out new deployments rapidly and regularly, which makes some errors like this unavoidable.
"We iterate faster every year, and as we have increased our iteration rate we realise that there constantly will be bugs," says Ukert. "It's about how quickly we can squash those bugs and have iterations. That's really what we find a monitoring tool helps us with."