From Two Decades of History Trapped in Expensive Servers and Databases

When a media company spends more than twenty years collecting data from every rugby match played, its most valuable asset stops being a website and becomes a unique historical data source. Thousands of results, statistics, photos, and records accumulated over two decades. Rugby Champagne had that history—in fact, the most complete rugby archive in the country—but it was locked inside a system that had become more of a liability than an advantage.

The Starting Point

The platform ran on traditional virtual machines and a SQL Server database. On paper, it worked (being generous with the definition of worked). In practice, it suffered from the typical limitations of an architecture designed for a different era:

  • High fixed costs: SQL Server licenses and always-on servers had to be paid for in full every month, whether there was traffic or not. The company was paying year-round for the capacity needed to survive the busiest day of the year.
  • No scalability: Whenever a major match or breaking news story drove a spike in traffic, the site would slow down or even go offline. Exactly when audience demand was highest—the moment a media outlet can least afford downtime.
  • A growth ceiling: Growth meant purchasing and configuring additional servers in advance. Infrastructure wasn't supporting the business; it was holding it back.

The underlying fear (the same one that stops most organizations in this situation) was simple: if we touch this system, will we lose twenty years of data?

The Decision

The premise was clear from the beginning: migration could not mean starting from scratch or sacrificing the historical archive. Everything published over the last two decades had to survive, remain accessible, and preserve its URLs to avoid broken links and lost search rankings.

That meant this wasn't a matter of "turning off the old system and turning on the new one." It was a phased migration, with old and new systems coexisting while data was transferred, validated, and only then retired once the replacement had been fully tested.

At no point did the website go offline.

The Solution

We replaced the always-on server model with a cloud-native architecture where every component has a single responsibility and costs are tied directly to actual usage:

  • The database was migrated from SQL Server to managed PostgreSQL, an open, high-performance solution with no licensing costs.
  • The application moved away from fixed virtual machines to serverless infrastructure that automatically scales up when traffic arrives and scales down when it doesn't.
  • Heavy content (images, files, galleries) was migrated to cloud object storage and delivered through a global distribution network.
    • For end users accessing the website and mobile app, an edge caching layer (CDN) was added. Pages are served from the location closest to the user, reducing load on the platform and improving performance across the board.

The result is a platform that automatically adapts to demand instead of forcing the business to predict and pay for future capacity in advance.

The Technical Details (For Those Interested)

The migration involved moving from SQL Server running on virtual machines to a stack built on managed PostgreSQL (Supabase), a Node.js API running on Google Cloud Run (serverless, scaling on demand and down to zero), an Astro frontend with server-side rendering, media storage on Cloudflare R2, and edge caching through Cloudflare, including protection rules against abusive automated traffic. Each component scales independently.

The Results

Improved speed and consistency for both internal teams managing the data and the users consuming it. This increased operational efficiency while delivering a smoother experience for customers.

Always online. The platform maintains 99% availability. Traffic spikes are no longer a concern: the system handles unexpected surges without any noticeable impact on site performance.

Lower costs—and, more importantly, room to grow. Database licensing fees and idle servers were eliminated, transforming infrastructure spending from a high fixed cost into a variable cost aligned with actual usage. But the most significant benefit goes beyond savings:

Growth no longer requires upfront infrastructure investment. The platform now supports the business instead of imposing limits on it.

The archive remained intact. More than 20 years of historical data were successfully migrated. Nothing was lost along the way.

The Takeaway for Organizations Considering the Same Move

Replacing a system that "still works" can be intimidating, especially when it carries years of history. But maintaining expensive, inflexible infrastructure also comes at a cost—the monthly expense itself, and the inability to scale when new opportunities arise.

The good news is that modernization doesn't require choosing between the past and the future. Organizations can move to modern, cost-effective, self-scaling technology without losing any of the value they've already built.

That's exactly what we achieved with Rugby Champagne, the most important rugby data archive in Argentina.