TicketSwap Tames Demand Ups and Downs with AWS

TicketSwap is one of the largest ticket exchange sites in Europe, and it sees extreme swings in customer activity depending on the calendar. It offers tickets to a variety of events, with the greatest number of transactions generated by the big European summer music festivals.

Event tickets usually go on sale well in advance. Some people might buy tickets not realizing that friends have done the same, leaving them with extra tickets they can’t use. Meanwhile, others might miss out on tickets but still really want to go. As a result, there’s a high demand for ticket swapping as the event gets nearer.

Based in the Netherlands and founded in 2012, TicketSwap briefly tried operating its own server but experienced availability problems and couldn’t scale to meet rapidly growing demand. It chose Amazon Web Services (AWS) several years ago, primarily based on AWS’s ability to provide instant scalability and host services in Europe where most of TicketSwap’s customers are located.

TicketSwap began by putting its database on Amazon Relational Database Service (Amazon RDS), which instantly relieved problems with management and maintenance. It then moved its website to AWS too. The company is now all-in, and it chooses applications based on their ability to run on AWS.

Because TicketSwap’s service sees massive changes in demand by season and even by the day, scalability and automation are critically important. To get the best value from AWS services, load balancing is also key.

During the autumn and winter, when there are fewer events, TicketSwap uses 10 nodes for Amazon Elastic Kubernetes Service (Amazon EKS) and Amazon Elastic Compute Cloud (Amazon EC2). But its service demands can be as much as 10 times higher during the festival season, which runs from April through September. Demand peaks can arrive extremely rapidly and be very short-lived—a nightmare for a traditional data center. For instance, when Ed Sheeran performed in Amsterdam, tickets for the concert went on sale on a Friday at 10 AM. Fifteen minutes later, every ticket had been sold. Eager fans piled onto TicketSwap to set ticket alerts, hoping for returns. Thanks to AWS, the company was able to instantly scale to 15 times normal capacity. A half-hour later, demand on the website was back to normal levels.

Pascal de Vink, lead architect at TicketSwap, explained how AWS helps cope with such wild fluctuations in demand: “The AWS Cost Explorer was a tremendous help in figuring out where we can optimize our spending. With the recent changes, this was an invaluable tool, just like AWS Budgets. In the end, what allowed us to improve our spending was being able to change the Amazon RDS and Amazon EC2 instance types without any noticeable downtime.”

TicketSwap’s relationship with AWS has evolved over time as its understanding of AWS services has matured and its needs have changed.

While it began by using Amazon EC2, the company has since improved resource usage with Amazon EKS and Amazon EC2 Spot Instances, which provide finer-grained control and help to reduce waste.

And Amazon CloudWatch provides insights to make sure systems run effectively. When TicketSwap customers experienced slow responses on the website, the company was quickly able to trace the problem to high CPU usage and then resolve the issue.

TicketSwap is now using infrastructure as code to manage all of its AWS services. This provides flexibility and scalability and makes it easy to add new domains or databases with a single click. It also provides transparency into the entire architecture without the need to scroll through multiple management dashboards. Increasingly, TicketSwap’s internal tools also rely on AWS and its APIs.

The COVID-19 pandemic put a stop to almost all large-scale events during the summer of 2020. However, despite this, TicketSwap is still seeing large numbers of visitors to its website.

With event activity dominated by smaller sporting events, theatrical productions, and day trips, people still have a need to swap or offload tickets. And AWS Auto Scaling means TicketSwap no longer needs to pay large sums for unused compute and storage capabilities.

TicketSwap is now planning for business in the post-COVID era. And it’s exploring other ways in which AWS services could help further improve its customer offering.

“In the short term, we are looking into the use of AWS Graviton Processor, to see if it can provide us with an even better cost versus performance tradeoff,” De Vink says. “In the longer term, we really believe that machine learning is a way to make life much easier for our customers. We already have a great track record when it comes to battling fraud, but maybe AWS Fraud Detector can enhance it even further. AWS Personalize might be interesting to us for recommending events that fans will love, with even more precision.”