Cutting the cord: migrating off a parent company before the clock runs out

Most migrations you choose. This one came with a contract and a countdown. After the platform changed hands, we had a transition-services agreement to exit: a fixed window to get ourselves off the former parent company’s systems and onto our own, after which the shared services simply went away. Miss the date and things stop working. There is no “we’ll finish it next sprint.”

The scope was everything you don’t think about until you have to leave: source code hosting, authentication and secrets, single sign-on, tooling, domains. All of it had quietly grown roots into the parent’s infrastructure over years, and now all of it had to come out without the platform skipping a beat.

The work is discovery before it’s migration

The thing nobody tells you about a divestiture cutover is that the migration is the easy half. The hard half is finding everything. Every pipeline that authenticated against the parent’s identity provider. Every repo that pulled config from a host you no longer control. Every integration that assumed an SSO that’s about to vanish. The coupling is invisible precisely because it always worked, so nobody documented it.

So we inventoried first. What authenticates against what. What pulls from where. Which domains resolve through whose infrastructure. Moving auth off the old GitHub and SSO onto our own Vault and a new identity provider wasn’t one task, it was a hundred small dependencies that each had to be found before any of them could be cut.

Cut over in reversible steps

With a hard deadline and zero downtime tolerance, the temptation is to rush a big-bang switch. That’s exactly backwards. The way through is reversible steps: stand up the new system alongside the old, prove it with a sandbox and a slice of real traffic, migrate one dependency at a time, and keep the old path until the new one is confirmed. A deadline doesn’t mean move recklessly. It means sequence ruthlessly, so that if any single step goes wrong it’s a small, recoverable problem and not a platform outage on a day you can’t afford one.

The takeaway

A divestiture migration combines the three things that make any change scary: a deadline you don’t control, coupling you can’t see, and a system you can’t take down. Respect that by spending your early time on discovery, not heroics, then cutting over in small reversible moves with the old path as your safety net until the last possible moment. The teams that exit cleanly aren’t the ones that moved fastest. They’re the ones that found everything first.