Fluid Studio: Building Scalable Website Infrastructure for the Gaming Industry on AWS

Overview

Fluid, a premier digital creative agency specializing in the gaming industry, partnered with Lambert Labs to architect a high-performance voting and awards platform for a major mobile gaming company. The project required an infrastructure capable of handling extreme, instantaneous peak traffic spikes the moment global voting was released. To meet these demands, the solution utilized a decoupled, cloud-native architecture on AWS featuring Amazon Elastic Container Service with AWS Fargate for independent autoscaling of web and worker layers. By integrating Amazon SQS to manage high-volume voting traffic asynchronously and Amazon Aurora Postgres for reliable, provisioned data storage, the platform successfully supported peak traffic spikes. This resilient AWS environment, supported by Amazon CloudFront for global content delivery and Amazon ElastiCache for session management, ensured a seamless user experience under the most demanding conditions while establishing a proven technical blueprint for future high-traffic digital events.

Working with Lambert Labs allowed us to launch a
high-stakes gaming awards
platform with absolute
confidence.
Lewis Bradburn, Digital Director, Fluid

Opportunity / Customer Challenge

Fluid, a digital creative agency, was commissioned by a major mobile gaming company to develop a voting and awards platform for a high-profile, one-off event. The site faced an immediate technical challenge: it had to handle extreme, instantaneous peak traffic the moment voting was released. To ensure zero downtime, Fluid needed an autoscaling solution that could process millions of votes within the first 24 hours without overwhelming the underlying database. The opportunity was to architect a resilient, cloud-native infrastructure on AWS using ECS on Fargate and SQS to decouple the high-volume voting traffic and ensure a successful global launch.

Solution

To handle the extreme peak traffic required by the Fluid project, Lambert Labs architected a highly scalable and resilient infrastructure centered on ECS on Fargate. The solution was designed with a decoupled architecture, separating the web layer from the worker layer to ensure that massive spikes in user activity would not disrupt the core voting process. To manage the influx of millions of votes, the voting infrastructure utilized SQS to buffer incoming requests, allowing the system to ingest data at high velocity while processing it asynchronously.  A critical component of the database management strategy involved the worker tier that processed votes from the SQS queues. To protect the Aurora Postgres database, which was provisioned rather than serverless to ensure consistent performance, the worker layer was intentionally capped at a maximum of two tasks. This intentional throttling ensured the database remained at a comfortable load, preventing failure during peak write activity while maintaining the integrity of the voting data. 

To optimize frontend performance, the architecture utilized an Elastic Load Balancing Application Load Balancer (ALB) and CloudFront distribution for caching static assets and server-side generated objects. While scaling based on CPU load was initially considered, load testing revealed it to be ineffective for this specific traffic profile; instead, the web and API servers were configured to autoscale based on the number of concurrent connections at the ALB. Additionally, ElastiCache Valkey was implemented to cache sessions and signups, further offloading the primary database and improving response times. 

Before the live launch, the entire infrastructure was subjected to rigorous load testing using Locust. These tests allowed the team to fine-tune the autoscaling parameters and confirm capacity limits, ensuring the system could handle the anticipated load.

Working with Lambert Labs allowed us to launch a high-stakes gaming awards platform with absolute confidence. Their expertise in architecting a decoupled, autoscaling solution was critical in handling millions of fans casting votes with zero downtime. The rigorous load testing they conducted ensured our infrastructure remained performant under extreme peak conditions, proving we can support extreme traffic at scale. (Lewis Bradburn, Digital Director, Fluid)

Outcome

The implementation of a decoupled, autoscaling infrastructure on AWS delivered an exceptionally successful launch for Fluid and their mobile gaming client. The platform seamlessly handled peak traffic with zero downtime. By utilizing a robust load testing strategy, Fluid was able to enter the launch window with complete confidence in their technical limits, ensuring the site remained performant even during the most intense traffic spikes.

This project highlights how a decoupled architecture using Amazon SQS can effectively manage extreme, write-heavy traffic spikes while protecting the core database. By pairing independent AWS Fargate autoscaling with rigorous load testing, we enabled Fluid to handle a high volume of users with zero downtime, proving that a resilient cloud-native design is essential for high-stakes digital events. (Rafiq Hilali, CTO, Lambert Labs)

About Fluid

Fluid (fluidstudio.co.uk) is a premier UK-based full service creative agency that specializes in crafting high-impact digital experiences, particularly within the gaming and entertainment sectors. They provide a comprehensive suite of services ranging from branding and UI/UX design to complex technical engineering for global mobile gaming brands. By bridging the gap between creative vision and robust technology, Fluid empowers its clients to launch large-scale campaigns and platforms that engage large numbers of users under high-pressure conditions. Renowned for their deep industry expertise, the agency has established itself as a trusted partner for brands seeking to deliver seamless, resilient, and visually stunning digital products.