AccessPilot Knowledge Base

Contents › Method

Reviewing emergency access (break-glass) accounts

A full checklist for the accounts that exist for the day everything else fails.
Published by Urbannerd Consulting · Reviewed 2026-07-29

Emergency access (break-glass) review

Scenario 6, and the gate on every other scenario. Sources: benchmark guidance / 2.2.1 (distilled in

) and Microsoft's own guidance

(). Where they differ, see

§5 — the checklist below is the

combined, stricter position.

The five scenarios these accounts exist for

Useful when a customer needs convincing. Microsoft names them:

  1. Accounts are federated and federation is unavailable — IdP outage or network break.
  2. Admins registered Entra MFA and all their devices, or the service, are unavailable

e.g. a carrier outage where phone call and SMS were the only registered methods.

  1. The last person with Global Administrator access leaves. Entra prevents deleting the

last GA in the cloud, but not deleting or disabling it on-premises.

  1. Natural disaster — mobile and other networks unavailable.
  2. **All GA and Privileged Role Administrator assignments are eligible, activation requires

approval, and no approvers are selected** (or the selected approvers left the directory).

Active GAs are the default approvers; with none active, **nobody can approve activation

and tenant administration is locked** — with no policy misconfiguration at all.

Scenario 5 is a PIM configuration trap. Test for it whenever a customer describes a

fully-eligible admin model.

Review checklist

Work every item. Report each as Pass / Fail / Not stated — "not stated" is a distinct

result and must never be silently scored as a pass.

Accounts

#CheckRequired stateSeverity if failed
1CountTwoCritical
2NamingDoes not identify any individualMedium
3DomainDefault .onmicrosoft.com, not a custom domainHigh
4Directory typeCloud-only, not synced from on-premisesCritical
5LicensingUnlicensedMedium
6Sign-in stateEnabled — not blocked, not disabledCritical
7RoleGlobal Administrator, permanently assignedCritical
8Role assignment stylePermanent, not PIM-eligibleCritical

On #3: custom domains can expire, be mis-transferred, or have DNS broken. The

.onmicrosoft.com domain cannot fail the same way, so the recovery account stays reachable

during exactly the failure it exists for.

On #8: a PIM-eligible break-glass account is not a break-glass account. If PIM or the

activation path is unavailable, the account cannot be used. benchmark guidance permits permanent

Assigned state specifically and only for up to two break-glass Global Administrators

(5.3.1).

Authentication

#CheckRequired stateSeverity if failed
9MFA presentYes — mandatory for all accounts since 15 Oct 2024, break-glass includedCritical
10Method typePasskey (FIDO2) or certificate-based authenticationHigh
11Method independenceDoes not depend on a phone, carrier, or the Authenticator push serviceHigh
12Password strength≥ 16 characters, randomly generatedHigh
13Password custodySplit into parts held by different people, or equivalent dual controlMedium
14Physical custodyFIDO2 keys in a secure, separate, fireproof locationMedium
14aMethod differs from normal admin accountsIf regular admins use Authenticator, break-glass uses FIDO2High
14bCredential/device does not expireNot subject to automated cleanup for lack of useHigh
14cNot tied to any employee-supplied deviceNo personal phonesHigh

On #11 — the failure mode people miss: if both accounts use Authenticator push and the

outage you are recovering from is an MFA service outage or a carrier failure, neither

account works. The methods must not share a dependency with the thing most likely to fail.

Conditional Access

#CheckRequired stateSeverity if failed
15Exclusion coverageAt least one account excluded from every CA policyCritical
16Exclusion mechanismDirect user exclusion, or a role-assignable / PIM-managed groupCritical
17New-policy disciplineA documented step requiring the exclusion on every new policyHigh
18Exclusion verificationConfirmed policy-by-policy, not assumedHigh

On #16 — the highest-value finding in this whole review. If the exclusion group is an

ordinary security group, anyone holding Group Administrator can add themselves to it

and bypass Conditional Access entirely, and can also remove the break-glass account without

detection. It must be a role-assignable group or managed through PIM for Groups (licensed).

On #15 — the trade-off to present honestly rather than resolve: benchmark guidance requires at least one

account excluded from all policies. Excluding both maximises recoverability but doubles the

unprotected surface. Excluding one keeps a hardened spare but risks the protected account

being caught by the same failure. State the trade-off and let the customer choose.

Monitoring

#CheckRequired stateSeverity if failed
19Sign-in alertingAlert on every sign-in, to a monitored destinationHigh
20Alert priorityHigh severity, treated as an incidentMedium
21Alert routingGoes to people who will actually see it out of hoursHigh
22Alert testedVerified by an actual test sign-inHigh
23Audit log alertingChanges to the accounts also alertMedium

Implementation options — benchmark guidance accepts any method that covers all break-glass activity:

severity, category Privileged accounts, single activity, filters Activity type = Log on

and User Name = <account>.

tenants without Defender for Cloud Apps. Requires an Azure subscription.

Do not present the Defender route as the only compliant one.

Governance

#CheckRequired stateSeverity if failed
24Written procedureAuthorised and distributed by senior managementHigh
25Usage authorisationDefined — who may authorise use, and whenMedium
26Test frequencyAt least every 90 days (Microsoft), plus a quarterly test that the accounts still sign in under current CA configHigh
26aRetest on trigger eventsOn IT staff change (termination or role change) and on subscription changeMedium
26bSecure workstationA PAW or designated secure client environment is used when accessing these accountsMedium
26cTwo independent network pathsIf a device is involved, it can reach the internet via two paths with no common failure mode (e.g. facility Wi-Fi and cellular)High
26dOn-prem / cloud separationEmergency access for on-premises and for cloud are distinct, with no dependency either wayHigh
27Last testDocumented, with the resultHigh
28Credential rotationAfter every use, and on a scheduleMedium
29Access reviewAccounts covered by a recurring access reviewMedium
30Restricted management AUAccounts placed in a restricted management administrative unitMedium
31Leaver processAnyone holding a credential fragment is covered by offboardingHigh

On #26 — untested break-glass is the most common real-world failure. Accounts sit for three

years, then during an actual incident the password is wrong, the FIDO2 key is

unregistered, the account was caught by a CA policy added last year, or the safe key is

missing. The test must be an actual successful sign-in, and it must confirm the alert

fired.

Output shape

  1. Summary — how many of the 31 checks pass, fail, or were not stated.
  2. Critical gaps — each with the mechanism by which it causes tenant loss.
  3. Full checklist table — Pass / Fail / Not stated, with a note per row.
  4. Remediation plan — in execution order.
  5. Test procedure — a concrete script for the next scheduled test.
  6. What needs tenant verification — the checks that cannot be answered in conversation.

Test procedure the agent should hand over

  1. Schedule the test and notify the security team, so the alert is expected.
  2. Sign in with break-glass account 1 from a clean browser session.
  3. Confirm sign-in succeeds and record the method used.
  4. Confirm the alert fired and reached the intended recipients.
  5. Confirm the sign-in log shows the account excluded from the CA policies, not merely

passing them.

  1. Perform a trivial, reversible privileged read to confirm the role is active.
  2. Sign out; clear the session.
  3. Repeat for account 2.
  4. Rotate credentials if the procedure requires it.
  5. Record: date, tester, result, issues found, remediation, next test date.

Common findings

FindingWhy it matters
Only one break-glass accountNo redundancy — one lost credential is total loss
Account on the custom domainDomain failure takes the recovery path with it
Account is PIM-eligibleUnusable if the activation path is what failed
Both accounts share an MFA dependencyBoth fail together in the outage they exist for
Exclusion by ordinary security groupGroup Administrators can bypass CA or silently remove the exclusion
Never testedDiscovered broken during the incident
Licensed with a mailboxUnnecessary attack surface
Alerting configured but never verifiedSilent failure — nobody learns until it matters
Password in the same vault as everything elseVault compromise takes the recovery path too
No documented authorisation procedureUse is indistinguishable from compromise