Client Infrastructure Handoff (Handoff)

Client Infrastructure Handoff (Handoff)

1. 01

   Freeze the delivery environment 48 hours before the handoff call

   Lock deploys, pause scheduled jobs, and tag the exact commit or container image the client is receiving so the handed-over state is reproducible.

2. 02

   Export every credential and secret into the client's own vault

   Repository tokens, SSH keys, database passwords, and DNS registrar logins move to client-owned storage; agency staff retain break-glass access only through a documented, time-boxed grant.

3. 03

   Map each pipeline stage to a named owner on the client side

   For a Git-triggered deploy chain, that means naming who approves a production push, who holds rollback authority, and who receives failure alerts at 2am.

4. 04

   Walk the client through one live deploy and one rollback

   Run both in front of them on a staging target. A platform with one-click rollback still fails a handoff if nobody on the client team has pressed the button.

5. 05

   Document the automation stack's known limits and exit paths

   Record where the current tooling stops scaling (multi-cloud provisioning, compliance-gated regions, data residency) and what migration would cost, so the client is not surprised later.

6. 06

   Transfer monitoring, alerting, and billing ownership

   Repoint uptime checks, log drains, and cloud spend to client accounts. Confirm the first invoice lands with them, not the agency, before sign-off.

7. 07

   Run a 14-day shadow period with a named escalation contact

   Agency stays reachable but does not act unless the client requests it. Log every request during the window; recurring ones become a retainer scope or a training gap.

8. 08

   Collect written acceptance against a defined checklist

   Sign-off should list what was transferred, what was explicitly excluded, and the date agency liability for production changes ends.