The Weekend Two Clients Called at Once.

The first phone rang at 9:14 on Saturday morning. The second rang less than two minutes later. Different clients. Different websites. The exact same problem. For a few moments, Fatou Ndiaye thought it had to be a coincidence. It wasn't.

As Technical Lead at Teranga Digital Studio in Dakar, Fatou wasn't responsible for a single product. Her team built products for everyone else — e-commerce websites, internal business tools, booking platforms, customer portals. If a client had an idea, her agency turned it into software. That meant living in several technical worlds at once.

Some projects ran on React. Others relied on Django and MySQL. Every client had chosen a different hosting provider over the years — different dashboards, different credentials, different invoices, different ways of deploying. Managing fourteen active projects often felt less like engineering and more like remembering where everything lived.

The work itself wasn't the hard part. The infrastructure around it was. Bringing a new client online routinely took three or four days — not because development moved slowly, but because every project required setting up yet another hosting account, configuring domains, requesting access from clients, exchanging credentials over email, and documenting everything before anyone dared push the first release.

"Every project solved a business problem. Every infrastructure setup created a new operational one."

That Saturday morning exposed exactly how fragile the system had become. A hosting provider suffered an outage, taking two unrelated client websites offline at the same time. Both companies called within minutes of each other, expecting immediate answers. The agency had no single place to understand what was happening — different providers meant different status pages, different support channels, and different dashboards to inspect before anyone could even confirm the issue.

Nobody panicked. But everyone realized they were spending too much energy managing hosting instead of serving clients. The following Monday, the team decided to rebuild their delivery process instead of simply recovering from another outage.

Every client project moved into a shared deployment workspace where applications could be managed from a single dashboard. Whether a project used React, Django, or MySQL no longer changed the operational workflow. New deployments became consistent, regardless of the underlying stack, and every engineer finally knew where to look before opening five browser tabs.

The agency reorganized access at the same time. Developers kept deployment permissions, project managers could monitor application health without risking accidental changes, and account managers finally had a reliable way to answer clients without interrupting engineers. Infrastructure stopped living inside individual laptops and became part of the team's shared process.

Even onboarding changed. New projects no longer started with creating hosting accounts across multiple providers. Databases could be provisioned immediately, custom domains connected with a single DNS change, and SSL certificates appeared automatically. Instead of treating infrastructure as the first week of every project, the team could focus on delivering value from day one.

Managing fourteen production environments no longer meant juggling six different hosting providers. It meant opening one dashboard and immediately seeing what was healthy, what was deploying, and what needed attention.