Eventbrite Simplifies Infrastructure, Reallocates Resources Using Amazon Aurora

Eventbrite is a global self-service ticketing company that connects event organizers to people around the world through live experiences. The company makes it simple for users to showcase, promote, and purchase tickets for live events. The business is dependent on quick and timely ticket sales, so limiting outages and maintaining website availability are top priorities to keep the company running and to earn customer trust. When the demand for live events decreased significantly during the COVID-19 pandemic, Eventbrite pivoted its focus from new product development to technological and foundational improvements in the business, using Amazon Web Services (AWS). The company wanted to create a robust and scalable infrastructure that would be able to successfully serve the anticipated increased demand for live events as the COVID-19 pandemic restrictions lifted.

Previously, the company was using a self-managed database solution that relied on a complex network of custom tools and required manual remediation when an instance went down, costing the company valuable time and potential lost revenue. To improve the reliability and reduce the complexity of its infrastructure, Eventbrite migrated from its self-managed MySQL database solution to Amazon Aurora—a MySQL- and PostgreSQL-compatible relational database built for the cloud that provides the performance and availability of commercial-grade databases at a fraction of the cost. As a result, Eventbrite improved its scalability and automation, reduced failover time, and saved on data storage costs. Additionally, the company reallocated resources to application teams, which work to deliver new features and functionality for end users and are proactively creating value for the core business.

Eventbrite integrates all the components event creators need to plan, promote, and produce live events so they can focus more effort on the event itself rather than the logistics.

Eventbrite started using AWS in 2014, but the company still had a complex network of self-managed MySQL databases. “We had a customized stack that only one or two people really knew inside and out,” says Chris Brown, engineering director at Eventbrite. “Our aging database infrastructure was becoming an increased liability to the business, especially from an operational support perspective.” The COVID-19 pandemic provided an opportunity for the company to refocus resources on infrastructure improvements as the demand for live events decreased. Eventbrite wanted to improve reliability and sustainability by migrating to an industry-standard database and decided to implement Aurora because of its benefits as a managed database-as-a-service solution and its compatibility with existing AWS services.

By the end of 2020, Eventbrite began migrating to Aurora MySQL. To minimize interruptions, it started by shifting 1–2 percent of its read traffic to Aurora to test performance and reliability in small increments. Over time, the company ramped up this percentage until all the read traffic was handled through Aurora and until it could migrate to the database-as-a-service solution for the writer nodes as well. “On the day of our migration, the AWS Infrastructure Event Management practice supported us through an all-hands-on-deck approach to monitor the cutover and make sure everything went smoothly,” says Vivek Sagi, chief technology officer at Eventbrite. “The network of partners, the processes, and the focus on our success at AWS were exactly what we needed to make this transition happen.” Because Eventbrite was significantly overhauling the database infrastructure that had been in place for nearly a decade, the company prepared its customers for potential downtime. In the end, however, this preparation wasn’t necessary. “We did not have any major issues the entire way,” says Brown. “Every one of the application migrations went exactly as planned, and we didn’t have any downtime.”

Using Aurora, Eventbrite saw immediate improvements in automation and read scalability. Issues that previously required management overhead and manual intervention could be handled by the fully managed service. “Previously, something that should have been a 30-second problem would be a 10-to-15-minute problem because someone had to be paged and manually make a configuration update,” says Brown. “Now we use what’s provided out of the box instead of rolling out homegrown solutions.” With those efficiencies, the company is able to automatically scale its resources, achieve high availability, and better prepare for the increased demand for in-person events as the world reopens from COVID-19 pandemic restrictions.

By simplifying its infrastructure, Eventbrite was able to reduce the number of engineers required to support its database by 75 percent. Additionally, Eventbrite’s hiring and onboarding process for new staff is simpler when using an industry-standard database instead of a custom solution. “Instead of taking months, we can onboard people who have experience with different AWS technologies in a week and a half,” says Brown.

Eventbrite also realized cost savings when migrating from its self-managed infrastructure to Aurora. Its previous database had approximately 50 database instances that replicated the data locally, each with its own storage layer. Using Aurora, the company set up database nodes that share the same underlying data store, which reduced data storage costs significantly.

Eventbrite plans to continue replacing its custom solutions company wide because of the immediate improvements it has seen in reduced costs and increased efficiencies from using fully managed AWS services. “Without a doubt, the biggest benefit of transitioning to Aurora is the reduced complexity in customization that we had built up over the years,” says Brown. “The simplicity directly leads to improved reliability and scalability.”