Supporting Chevening through its highest-traffic season, every year

One yearly peak, zero room for error

Chevening's site carries country pages, blogs, events and application forms for a global audience of scholars and fellows across 100+ countries. Once a year, applications season drives an enormous, concentrated spike in traffic and form submissions — and the existing single-server setup had no headroom for it, leaving the team exposed to slowdowns and outages at the exact moment they could least afford them.

Built to scale before the season starts

• Ran a full architecture review of the existing AWS setup: single points of failure, database contention under load, and where the current stack would break first under peak traffic.
• Produced a high-level design mapping the target architecture — load-balanced application tier, managed database tier, autoscaling policy — then a low-level design specifying instance sizing, RDS configuration, failover behaviour and monitoring thresholds.
• Rebuilt the hosting on managed, load-balanced servers behind an autoscaling application tier, with the database migrated onto RDS for managed failover, automated backups and read replicas.
• Put the whole estate on our standard Drupal support SLA — security patching, 24/7 monitoring and proactive capacity planning ahead of each application window.

99.99% uptime, even at peak

• 99.99% uptime maintained through applications season, including its highest single-day traffic on record.
• Average load time dropped to 2.1s (from 6.4s) after the architecture rebuild and image optimisation.
• Accessibility score rose to 96, keeping the site aligned with public-sector accessibility requirements.
• The application tier now scales automatically ahead of demand instead of the team having to pre-provision capacity by hand each year.