Autoscaling a high traffic events platform with Students' Union Management System (SUMS)

Overview

Students’ Union Management System (SUMS), a specialist UK-based Students’ Union system with an event ticketing component, struggled with website slowdowns and failed purchases during peak traffic spikes. To improve scalability and performance, the system was migrated from AWS Elastic Beanstalk to Amazon Elastic Container Service (Amazon ECS) with AWS Fargate, enabling faster and more efficient scaling. This transition ensured a smoother ticketing experience during high-demand periods while reducing operational overhead, allowing SUMS to focus on enhancing its platform for students’ unions.

Lambert Labs turned our vision into reality which has
dramatically improved our product’s performance.
Alan Bage, Principal Solutions Architect, SUMS

Opportunity / Customer Challenge

SUMS is a UK-based web platform that provides specialist features to run Students’ Unions efficiently, including selling event tickets online. By providing a seamless way for students to book and manage event attendance, SUMS plays a vital role in student life across universities.

As demand for student events grew, SUMS faced a major challenge: when tickets went on sale, sudden spikes in website traffic caused slowdowns, errors, and failed purchases. Their existing system wasn’t responding quickly enough to these surges, leading to long wait times and frustrated users.

While some high-traffic periods could be predicted, many were unexpected, as students’ unions scheduled events independently—sometimes at the last minute. This made it difficult to prepare in advance, and scaling the system too early or too late led to unnecessary costs or poor performance. To continue providing a reliable ticketing experience, SUMS needed a solution that could handle sudden traffic spikes quickly while keeping costs under control.

Solution

As containers scale faster than virtual machines, the application was moved from AWS Elastic Beanstalk to AWS Fargate. To test and quantify autoscaling performance under real world conditions, a distributed load testing solution from the AWS Solutions Library was deployed. Using this solution, load testing was performed against a full demo version of the SUMS application (a 3-tier web application using Elastic Load Balancing, AWS Fargate with Application Auto Scaling, Amazon Relational Database Service (Amazon RDS) and AWS Secrets Manager for credentials management).

As the initial results were disappointing (autoscaling performance not significantly better than that already experienced in Elastic Beanstalk, e.g. ~3 minute wait between initial spike and autoscaling), the default Central Processing Unit (CPU) metric used for autoscaling was changed to a custom metric in Amazon CloudWatch. The application container was modified to send CPU usage every 10 seconds to the custom metric (compared to the default metric’s every 60 seconds). This yielded much better results, e.g. a less than one minute wait between initial spike and autoscaling. To improve robustness, a dedicated process manager (supervisor) was used to manage the sending of the custom metric.

We knew we wanted to migrate from Elastic Beanstalk to Amazon ECS and use Infrastructure as Code (IaC), and that by doing so we could solve our peak traffic scaling issues. Lambert Labs turned our vision into reality which has dramatically improved our product’s performance during ticket releases. (Alan Bage, Principal Solutions Architect, SUMS)

Outcome

By adopting AWS Fargate with custom Amazon CloudWatch metrics for autoscaling, SUMS significantly improved its ability to handle high-traffic ticket sales. The time to scale new containers / instances in response to demand was reduced from several minutes to under one minute, ensuring a smoother experience for users during peak sales periods.

The implementation of a dedicated load testing environment enabled SUMS to validate its new infrastructure under real-world conditions, helping the team fine-tune scaling parameters and optimise system performance. The new architecture also reduced operational overhead by eliminating the need to manage Amazon Elastic Compute Cloud (Amazon EC2) instances, allowing SUMS to focus more on improving its platform rather than infrastructure maintenance.

With a more responsive and cost-efficient system in place, SUMS can now support student unions with confidence, knowing that its platform will scale seamlessly to meet demand—whether expected or sudden.

Scaling web servers in seconds in response to sudden traffic spikes is rarely straightforward. We were pleased to see Amazon ECS combined with Application Auto Scaling and Amazon CloudWatch alarms handled extremely high and rapidly increasing request volumes without issues. (George Lambert, Founder & CEO, Lambert Labs)

About Students’ Union Management Systems

Students’ Union Management System (SUMS) is more than just software – it’s a purpose‑built partner for Students’ Unions. Designed specifically for the sector, SUMS brings together membership management, events, insights, and engagement into one seamless, secure platform that grows with your union. By streamlining processes and reducing admin, it frees your team to focus on what matters most: strengthening community, empowering student leaders, and delivering an outstanding experience for every student. Built on collaboration rather than profit‑first thinking, SUMS provides the tools, data, and support unions need to evidence their impact and make confident, strategic decisions that shape thriving student communities.