When the one admin who holds every password is leaving
A step-by-step handover plan for when your only IT administrator is leaving, including break-glass accounts, domain ownership, MFA devices and the hostile-exit sequence.
What happened at an Army contractor
In December 2016, a system administrator at Federated IT, a contractor running servers for the US Army Chaplain Corps, learned his job was ending. According to reporting on the federal case, he removed every administrator account except his own, had the domain’s DNS registration moved into his name, and deleted project files that included network diagrams and the login details for the cloud accounts. The result was an outage of the Army chaplains’ support portal: pastoral staff could not sign in and students lost access to training material. After he was fired, he launched denial-of-service attacks against the servers in the weeks that followed. He pleaded guilty to unauthorized access to a protected computer and in 2019 was sentenced to two years in prison and ordered to pay about $49,000 in restitution.
That is an extreme case. The everyday version has no criminal in it. A loyal administrator retires or takes another job, and the organization finds out the domain is registered to his personal email, the backup console is tied to his phone, and nobody else knows the password manager’s master password. The cause is the same in both stories: one person holds all the access.
Why this is a leadership problem
CISA defines an insider as anyone who has or had authorized access to an organization’s resources, and it lists sabotage of physical or virtual infrastructure as one way an insider threat can show up. CISA also names termination as a possible motive. That definition includes your most trusted person. The fix is not about suspicion. It is about making sure no single departure, friendly or not, can lock you out of your own business.
Microsoft says the same thing in its emergency access guidance for Entra ID. One of the situations it plans for is the last person with Global Administrator access leaving the organization. Microsoft Entra ID will not let you delete the last Global Administrator account, but it does not stop that account from being deleted or disabled on-premises, and either one can leave you unable to recover.
The handover plan, step by step
Start as soon as you know about the departure. A two-week notice period is enough if you work in this order.
1. Build the inventory before anything else
Sit down with the departing administrator and list every place that has an administrator or owner role. Do not rely on memory alone. Work through invoices, the password manager’s vault and the email inbox to find services. At minimum:
- Identity: Microsoft 365 or Google Workspace super-admins, on-premises Active Directory domain admins
- Domain registrar account and DNS host, which are often two different companies
- Firewall, Wi-Fi controller, VPN and remote access tools
- Backup console and any offsite or cloud backup account
- Line-of-business applications, payroll, banking portals with admin users
- Cloud accounts (Azure, AWS, hosting) and their billing owners
- Vendor and MSP portals, software license accounts, the phone system
For each, record who the account belongs to, which email address recovery messages go to, and which MFA device approves sign-ins.
2. Create break-glass accounts that belong to the organization
Microsoft’s guidance is a good model even if you do not use Microsoft. It recommends two or more emergency access accounts that are cloud-only (on the onmicrosoft.com domain, not synced or federated), are not tied to any individual user, and do not depend on anyone’s personal phone. They should use phishing-resistant sign-in, with a FIDO2 passkey recommended, and that method should differ from what your normal admin accounts use. Microsoft also says to store the credentials in secure, fireproof safes in separate locations, exclude these accounts from Conditional Access policies that block sign-in, alert on every sign-in, and test the accounts at least every 90 days and whenever IT staff change.
Set these up while the departing administrator is still available to help, and have a second person, usually the owner or a senior manager, complete the first test sign-in.
3. Fix domain and DNS ownership
Your domain is how email, your website and many password resets work. ICANN notes that paying for a domain is not the same as owning it outright, and that a developer or other third party may have registered it with their own contact details, making them the registrant of record. ICANN’s good-practice guidance says the organization’s legal name should be the registrant organization, the registrant name should be a role such as a department rather than an individual employee, and outside parties should not be the registrant.
Log in to the registrar together, confirm the registrant is the company, move account recovery to a role mailbox the business controls, add a second authorized user, and do the same at the DNS host. A change to registrant contact information can also trigger a transfer lock at many registrars (ICANN’s page describes a 60-day lock with an opt-out), so check your registrar’s current rule and make the contact changes now, not on the day you need to move the domain.
4. Move MFA off personal devices
Go through the inventory and look for any admin login whose second factor lives on the administrator’s personal phone. Re-enroll each one with a method the organization controls: a company-owned security key kept in the safe, or a company-managed authenticator. Do this one service at a time, with the administrator present, and test each sign-in before moving on.
5. Hand over the password manager properly
If the vault sits in the administrator’s personal account, export only the business entries into an organization-owned business vault, with at least two people holding admin rights. Then treat every credential that person knew as exposed and rotate it on a schedule: privileged accounts first, then shared service accounts, then everything else. Rotation is routine hygiene, not an accusation, and you should say so.
6. Revoke on the last day, then confirm
On the final day, disable the person’s own accounts, remove them from every admin role on the inventory, revoke remote access and VPN certificates, and collect company devices and security keys. A week later, review sign-in logs for any of their old accounts.
The hostile-departure variant: preserve, then revoke
If the departure is a termination, or you have reason to think the person is unhappy, the order changes. You cannot spend two weeks working side by side.
| Step | What to do | Who |
|---|---|---|
| 1. Preserve | Before the conversation, export or snapshot audit logs, confirm a current backup exists and is stored where the admin cannot reach it, and record the current admin role list | Owner with a trusted outside party |
| 2. Secure the keys | Confirm the break-glass accounts work and the registrar login is in company hands. If they are not, get them set up first | Owner, outside IT help |
| 3. Revoke at the meeting | Disable accounts and remove admin roles at the moment the meeting begins, not afterward | Outside IT help, on standby |
| 4. Rotate | Change every shared and privileged credential the person knew, starting with identity, backups and remote access | Outside IT help |
| 5. Watch | Alert on sign-ins to old accounts and on changes to DNS, admin roles and backup settings | Whoever monitors your systems |
| 6. Document | Keep a dated record of what was done and when, in case you need it for an insurer, a lawyer or law enforcement | Owner |
Preservation comes first because revoking access does not undo anything that already happened. If damage has occurred, you need the logs and backups to show what changed and to restore. Talk to your employment lawyer about the termination itself, and if you suspect sabotage, about reporting it.
Make sure it does not happen again
The real fix is a short document and a schedule. Keep the privileged account inventory in a place at least two people can reach. Name an owner for each row. Keep break-glass credentials in the safe and test them quarterly, which also meets Microsoft’s 90-day floor. Put the registrar and DNS logins in the business vault. Review the list whenever someone with admin rights joins, leaves or changes jobs, and change safe combinations when someone with access leaves, as Microsoft recommends. Then a resignation letter is just a staffing problem, not a crisis.
When to get help
If you are not sure what your administrator controls, or the departure is already in progress, an outside review of privileged access and a planned handover is worth doing before the last day, not after.
Want to talk it through? Book a free 30-minute call with Security Medic Consulting.
Sources
- The Register, Holy smokes. Ex-IT admin gets two years prison for trashing Army chaplains’ servers, September 30, 2019 (accessed 2026-10-29).
- US Attorney’s Office, Eastern District of Virginia, Man sentenced to prison for cyber sabotage, September 27, 2019 (accessed 2026-10-29).
- CISA, Defining Insider Threats (accessed 2026-10-29).
- CISA, Insider Threat Mitigation (accessed 2026-10-29).
- Microsoft Learn, Manage emergency access admin accounts, June 4, 2026 (accessed 2026-10-29).
- ICANN, General Questions (FAQs for registrants) (accessed 2026-10-29).
- ICANN, Good Practices for the Registration and Administration of Domain Name Portfolios (Part II), June 23, 2017 (accessed 2026-10-29).
- ICANN, 5 Things Every Domain Name Registrant Should Know About ICANN’s Transfer Policy (accessed 2026-10-29).