Since launching in 2009, Grindr has been a trailblazer in the world of mobile dating. The first and largest gay dating network in the world, it now boasts two million daily active users in 198 countries. That means a lot of dates, and a lot of men looking for cool places to meet up.
But where should they go? Which nearby restaurants are hot right now? Which bar on this street serves the best martini? What awesome events are happening downtown tonight?
Grindr wants to answer those questions. Having mastered the art of connecting gay and bisexual men with one another, the company now aims to provide them with targeted information about entertainment and events of interest happening in their vicinity. In other words, Grindr is set to cement its position as a comprehensive lifestyle resource.
“Our ability to connect our users with a curated range of interests follows the natural evolution of our community’s high expectations,” explains Joel Simkhai, CEO of Grindr. “We’re very excited to utilize the geospatial technologies and machine learning that we have developed to enable a more holistic experience for our users.”
Grindr invests in scalability
Those users, a million of whom are active at any given minute, exchange 85 million chat messages and 3 million chat images each day. That adds up to almost a billion API calls. These spectacular stats provide a glimpse of the scaling challenges the company has faced—challenges now intensified by the evolution from hook-up app to lifestyle brand.
To meet them, Chief Technology Officer Lukas Sliwka implemented big changes to the Grindr stack when he joined the company in 2013. His first priority was to replace custom, home-built components with superior out-of-the-box solutions. Those solutions—including Elasticsearch, Redis, Akka and Java, and Apache Zookeeper—were easier to manage, he says, and enhanced Grindr’s scalability.
“I don’t believe in reinventing the wheel in terms of technologies that already exist,” says Lukas. “We form a lot of partnerships with companies that can enable my engineering organization to do things faster and better.”
Big stack, small teams
The Grindr stack now pulls together a wide array of cloud and open source technologies. Amazon Web Services provides scalable, cost-effective infrastructure. For backend messaging, Lukas and his colleagues rely on CloudAMQP’s RabbitMQ; for latency optimization, Cloudflare. And with Grindr’s caching needs continuing to grow, infrastructure from Redis Labs supports up to 400,000 operations per second.
Making sure all of these different tools work in harmony are more than 40 Grindr engineers. “We try to form small teams that are almost treated like mini startups within the company,” says Lukas. “Teams that are focused on a very specific KPI or business aspect.”
His role, and that of Grindr’s engineering managers, is to reduce friction as much as possible so that the engineers can execute efficiently. Each team determines the mode of agility that works best in terms of meeting its particular goals. “Our DevOps teams, for instance, are measured by their ability to resolve problems,” he explains.
Of course, the first step in resolving problems is identifying them. That’s why, early in 2016, Grindr added New Relic to its extensive toolkit.
Piece of cake
The metrics Grindr received from its performance monitoring tools before adding New Relic to the stack covered only the company’s own technology stack, Lukas says. “If I looked at the different APIs or business transactions, I wasn’t able to actually click on them and drill into what was really going on.” In other words, those tools didn’t provide sufficient insight into true app performance as actually experienced by Grindr users. “New Relic augments what we already had with its ability to explode each of those lines into a much more detailed view.”
Lukas is especially excited by the “layer cake” of visual data that New Relic creates for each transaction, allowing engineers to quickly pinpoint any problematic areas in the infrastructure or code, and to reduce their troubleshooting time, even in the most severe cases, to less than 30 minutes.
Currently using New Relic in development, Grindr is working towards implementing New Relic in production as well, giving access to some 25 engineers across DevOps and service reliability engineer (SRE) teams.
“Profiling your code in a very isolated environment is one thing, but being able to actually visualize what’s going on in real time in production—that’s another,” Lukas says. “That’s a very powerful offering that we are hoping to leverage.”
Join the evolution
Given Grindr’s fame for being not just mobile first but mobile only, Lukas is looking forward to the potential for frontend optimization offered by New Relic Mobile. “If I can trace what’s happening on the client in terms of network transit all the way down to the actual service-side execution, that will be very valuable.”
He’s right to be excited—and Grindr users should be, too. The app is set to become even more indispensable, and users can be part of its evolution. “Authenticity is crucial to our success, so we actively open dialog with our users to help shape the brand,” says Joel. “Bringing our users along for the journey is a big focus.”