QUEST Alliance is a not-for-profit focused on research-led innovation in teaching and learning. Its digital learning platform, the Quest App available on web and mobile, and supported by Accenture helps young people across the country build 21st-century employability skills, with the goal of reaching 1 million learners and helping them develop the skills, critical awareness, and confidence to seize opportunities and improve their communities.
With a major onboarding push of 250,000 users on the horizon, QUEST needed to be certain the platform could withstand the load. The catch: the existing application could handle only about 200 concurrent users and needed to scale to at least 10,000 - a 50x jump.
Scaling an application 50x isn’t about adding servers - it’s about finding what breaks first. The Quest App was about to go from a few hundred concurrent users to tens of thousands, and the existing architecture wasn’t built for it.
The application could handle only ~200 concurrent users but had to support at least 10,000, ahead of onboarding 250,000 new users.
The existing system broke at higher loads for reasons that first had to be diagnosed across the database, web-server configuration, and infrastructure.
Scaling couldn’t come at the cost of downtime; releases needed to be safe and low-risk.
Route 53 and Cloudflare introduced DNS issues, and Cloudflare caused load-testing timeouts at the 50x target.
Focaloid started with a technical analysis to find the bottlenecks, then implemented a targeted set of infrastructure, database, and web-server changes to carry the platform to 10,000+ concurrent users.
Ran a technical analysis of the existing system to identify exactly what was causing it to break under load.
Upgraded the Amazon RDS instance from R5.Large to R5.Xlarge and updated database indexing for query performance.
Implemented horizontal auto-scaling (2 → 4 medium servers based on load) behind an Application Load Balancer, with round-robin instance switching and sticky sessions enabled.
Updated mpm_prefork.conf and MaxRequestWorkers across all servers to handle far more concurrent requests.
Moved uploaded and generated files to Amazon S3.
Focaloid worked diagnosis-first, then changed only what the load demanded, and validated everything under stress.
Diagnosed the bottlenecks causing failures at higher loads, across database, web servers, and infrastructure.
Upgraded the database, added horizontal auto-scaling and load balancing, tuned web-server concurrency, and offloaded files to S3.
Implemented Blue-Green deployment via ALB to cut downtime and release risk.
Ran load testing to the 50x target, identified and fixed the Cloudflare-related DNS and timeout issues, and validated capacity.
For a non-profit working to put employability skills in the hands of a million young people, a platform that buckles at a few hundred users is a hard ceiling on impact. Scaling it 50x - safely, and without simply throwing money at always-on servers is what turns an ambition into a reachable goal. By diagnosing the real bottlenecks and re-architecting for elastic scale, tuned concurrency, and zero-downtime releases, Focaloid took the Quest App from ~200 to 10,000+ concurrent users while keeping cloud costs in check. That’s headroom the mission can actually grow into.
We diagnose the real bottlenecks and re-architect for elastic scale auto-scaling, load balancing, database tuning, and zero-downtime Blue-Green deployments - so you can grow without breaking or overspending.