The First Week After a Developer Goes Quiet

Accounts First, Code Later

Day one, the instinct is to open the code. Code is rarely what takes a service down. The domain registrar, the host, the payment processor, the transactional email service, every API key: start by finding out whose name each of those is in and which account holds it. Anything registered under the departed developer’s personal login takes the access away with them. Nothing looks wrong at first, because the site keeps serving. Then a renewal date passes and the domain drops or the server stops.

Is the Repository Running What the Server Is Running

Accounts secured, the repository comes next. Knowing the URL is not the same as having access, and having access is not proof that the code in there is the code currently serving traffic. Urgent fixes get made on the server at midnight and never pushed back. A new developer who treats the repository as the source of truth will wipe those out without noticing. Diff what is deployed against what is committed once, early, and that whole class of accident goes away.

The Config Is Not in the Repository

This is where a complete code handover still fails to run. API keys, database credentials, payment integration values: the sensitive settings usually live on the server or in a separate secrets tool, never in the repository. Without them a new developer cannot get the service up locally, let alone deploy it. If the previous developer is still answering messages, that is the thing to ask for, where the values are stored and who holds access to that store right now.

Expect No Handover Document

Sometimes there is documentation. More often there is not. Chasing someone who has already left for a document written after the fact is worse odds than having the incoming developer read through the code and the deploy flow and build their own map. Opening someone else’s codebase and getting the shape of it takes time, and that is normal rather than a sign of a slow hire. It is also why the opening stretch of a takeover produces questions instead of visible progress.

What Has to Happen This Week

Straight after a departure, sort by what can break. An expiring domain or an expiring server stops everything if it is missed, while a rough feature can sit untouched for a month without consequence. Give the first few days to expiry dates and access, the items that are painful to reverse once they blow, and read the codebase after that. Do it the other way around and a renewal date slips past while someone is halfway through the code.

Send Us What You Were Left With

Tracking down accounts one at a time and reassembling scattered config eats hours. Weple’s site care service takes over services in exactly this state and works through the expiry dates and the access list before anything else. Whatever is in your hands right now, a login that still works, a URL you remember, a file someone sent you once, forward it as is. Working out what is urgent is our part.