About this article: This article is reviewed against the cited sources. This is source research, not lab validation. Check the stated versions and your own requirements before making changes.
"Pending" is a state of the device object in Microsoft Entra ID. It means the on-premises computer account has been synced to the cloud, but the computer itself never finished registering. A pending device cannot do the things hybrid join is for: Microsoft's page says it "can't complete any authorization or authentication requests, such as requesting a Primary Refresh token for single sign-on, or applying device-based Conditional Access policies".
To find out why, run dsregcmd /status in an elevated Command Prompt on the device and read the Diagnostic Data section: the Error Phase and Client ErrorCode say where the join stopped.
What the error means
Microsoft's pending-devices page says the Microsoft Entra hybrid join task in Microsoft Entra Connect Sync syncs device objects to Microsoft Entra ID and temporarily sets their registered state to pending until the device completes registration. The state exists because the device must be added to the directory before it can register. The page adds that the pending state exists only for Microsoft Entra hybrid joined devices.
A device can get stuck in two ways, according to that page:
- A new on-premises, domain-joined device syncs to Entra ID but cannot complete registration, for example because it cannot reach the registration service.
- A device that was registered changes to pending. The page's scenario: the device object is moved to an organisational unit (OU) outside the Entra Connect sync scope. Sync treats that as a deletion and deletes the cloud device. When the object is moved back into scope, sync creates a pending device, and the device fails to register because it was registered before.
Microsoft's dsregcmd page does not describe a Pending value. The command shows the join state of the device itself. On a healthy hybrid joined device the page's expected values are DomainJoined YES, WorkplaceJoined NO and AzureAdJoined YES. AzureAdJoined NO means the join to Entra ID has not finished.
When the device is domain-joined but cannot hybrid join, the output includes a Diagnostic Data section with tests and a "Previous Registration" record of the last failed attempt. This appears on Windows 10 version 1803 or later.
Likely causes
| Cause | What you see | How to confirm |
|---|---|---|
| No line of sight to a domain controller (pre-check phase) | Error Phase pre-check; AD Connectivity Test FAIL | Device is off the internal network or a VPN with domain controller access |
| Service connection point missing, wrong or unreadable (discover phase) | AD Configuration Test FAIL; error 0x801c001d, DSREG_AUTOJOIN_ADCONFIG_READ_FAILED | The SCP in the forest points to a verified domain in the correct tenant. A second documented cause of 0x801c001d: event 220 in the User Device Registration log, where Windows cannot access the computer object in Active Directory; the page ties error codes 1312 and 1317 to replication issues and says to troubleshoot Active Directory replication |
Device cannot reach https://enterpriseregistration.windows.net or https://login.microsoftonline.com in the system context |
DRS Discovery Test FAIL; errors 0x801c0021, 0x801c001f, 0x801c003d, or network codes 0x80072efd, 0x80072ee2 | Test from the system context; a proxy that needs authentication blocks it |
| Proxy returns a page instead of data | Error 0x8007000d, E_INVALIDDATA | The page says the proxy is likely returning an HTTP 200 with an HTML authorisation page |
| Wrong tenant ID in the service connection point | 0x801c003a, DSREG_DISCOVERY_TENANT_NOT_FOUND; AADSTS90002 | The tenant ID in the SCP |
| Sync has not finished | Join error "The device object by the given id ... isn't found"; Client ErrorCode 0x801c03f2 | The page says this is expected for sync-join until Entra Connect Sync completes |
| Device moved out of the sync OU and back | Pending device that never registers | OU history; the device was registered before |
| TPM locked out | NTE_AUTHENTICATION_IGNORED (0x80090031) | Transient; the page says wait for the cool-down period |
| Throttled | "Your request is throttled temporarily. Please try after 300 seconds." | Expected after several registrations in quick succession |
Prove the cause before changing anything
All of these only read.
- On the device, open an elevated Command Prompt and run
dsregcmd /status. The dsregcmd page says to run it elevated so diagnostics run in SYSTEM context, which is closest to how the join runs. Copy the output into a text file. - In the Device State section, read DomainJoined and AzureAdJoined against the expected values above. WorkplaceJoined sits in the User State section, which the dsregcmd page says must be read from a user context, so read it in the affected user's own non-elevated prompt.
- In Diagnostic Data, read the tests: AD Connectivity Test, AD Configuration Test, DRS Discovery Test, DRS Connectivity Test, Token acquisition Test and Fallback to Sync-Join. Then read Previous Registration, Error Phase, Client ErrorCode, Server ErrorCode, Server Message and Https Status.
- Match Error Phase to the page's phases: pre-check, discover, auth and join.
- For earlier Windows 10 versions, use Event Viewer > Applications and Services Log > Microsoft > Windows > User Device Registration and look for events 304, 305 and 307 (phase and code), 201 (discovery) and 204 (join).
- For the moved-device cause, establish it before any fix: ask the owner of Active Directory whether this computer account was ever moved to an OU outside the Entra Connect sync scope and back, because the page gives no tool for reading that history. Also confirm that
dsregcmd /statuson the device shows DomainJoined YES and AzureAdJoined NO, and that the device is listed as Pending by the commands in the next step. Without all of these, treat the cause as unproven. - To list pending devices in the tenant, install the Microsoft Graph PowerShell SDK and sign in with
Connect-MgGraph. The page gives no scopes, so use an account that has permission to read devices, with scopes as your administrator directs. Count them:
(Get-MgDevice -All -Filter "TrustType eq 'ServerAd'" | Where-Object{((-not $_.AlternativeSecurityIds))}).count
The page's second form saves the pending devices, with their DisplayName and DeviceId, to a CSV file, so one device can be looked up by name. The command writes pendingdevicelist-summary.csv in the current folder and replaces a file of that name without asking, so run it in an empty folder:
Get-MgDevice -All -Filter "TrustType eq 'ServerAd'" | Where-Object{(-not $_.AlternativeSecurityIds)} | select-object -Property AccountEnabled, Id, DeviceId, DisplayName, OperatingSystem, OperatingSystemVersion, TrustType | export-csv pendingdevicelist-summary.csv -NoTypeInformation
Example (illustrative): at Contoso (contoso.com.au), laptop pc042 is domain-joined but shows as Pending. dsregcmd /status shows DomainJoined YES and AzureAdJoined NO. The Diagnostic Data shows AD Connectivity Test PASS, DRS Discovery Test PASS and a Previous Registration of sync with Error Phase join, Client ErrorCode 0x801c03f2 and "The device object by the given id ... isn't found". That is the expected sync-join message, so Contoso waits for Entra Connect Sync to finish, then retries. If the same laptop had been moved to a staging OU outside the sync scope and back, the cause would instead be the move.
Fix
Fix the proven cause. Do not apply every row.
- Network or proxy: Save first: export the current rules. Risk: the proxy and firewall changes below affect other traffic. Make sure the device can reach
https://enterpriseregistration.windows.netandhttps://login.microsoftonline.comin the system context, and that any outbound proxy accepts the computer account silently. To undo it: restore them. - Service connection point: Save first: record its current values. Risk: the SCP is read by every device in the forest, so the correction below reaches all of them. Correct it so it points to a verified domain and the right tenant. To undo it: put the values back.
- Sync scope: Save first: record the current OU selection. Risk: taking an OU out of scope deletes its cloud device objects, and putting it back creates Pending ones. Make sure the OU holding the computer accounts is in the Entra Connect sync scope, and that default device attributes are not excluded. To undo it: restore the selection, but a cloud device that sync has already deleted comes back only as a new Pending device; for a computer that registered before, that is the moved-device case below, and it needs the
dsregcmd /leavestep. For that reason only add OUs to the scope here; do not take one out to test. - Moved device that registered before: unregister it and let it start again, but only when step 6 of the proof list established that this device was moved outside the sync scope and back. A device with a network or service connection point fault would be unregistered for nothing.
Disruptive step: dsregcmd /leave followed by a restart, for the proven moved-device case only. Run it at an elevated command prompt, then restart the device. This unregisters the device from Microsoft Entra ID, so device-based Conditional Access and single sign-on stop working for it until registration completes again. Save first: copy the full dsregcmd /status output and note the device name and object ID in Entra. The page documents no way to restore the old registration. The supported route is that the device reinitiates registration through the scheduled task, which for Windows 10 devices is under Task Scheduler Library > Microsoft > Windows > Workplace Join > Automatic-Device-Join Task. Back-out: none; if registration does not return, read dsregcmd /status again and use the table above.
Do not change the FallbackToSyncJoin registry value to make an error disappear. The page describes it as the setting that controls whether sync-join is tried after an authentication failure.
Confirm the fix
- Run
dsregcmd /statusagain in an elevated Command Prompt for the device's join state. In Device State, AzureAdJoined should be YES and DomainJoined YES. - For the SSO State section, the page says to open the Command Prompt in the context of the logged-in user, so have the affected user run
dsregcmd /statusin their own non-elevated prompt. AzureAdPrt should be YES. If AzureAdPrtUpdateTime is more than four hours old, the page says to lock and unlock the device to force a refresh and check the time updates. - Check this one device is no longer Pending: run the CSV form of the command above and look for its DisplayName, or look at the device in the Entra admin center. The count command is tenant-wide and cannot show one device.
- Sign in as a user and test the device-based policy or single sign-on you were trying to fix.
What to watch afterwards
- The number of pending devices after each batch of new computers or OU moves.
- Computers moved between OUs. Moves outside the sync scope delete the cloud device.
- Devices that stay in the sync-join expected state for a long time; the page says the next join after sync completes should resolve it.
- Proxy changes on the network that carries system-context traffic.
Read next
Related articles you may want next.
- Troubleshooting"An Active Directory Domain Controller for the domain could not be contacted": fixing domain join DNSA new server will not join the domain: error 0x54b, 0x0000232A or 0xa8b. Read the Details text, prove whether DNS, the domain name you typed or a blocked port is the cause, and fix it.Read article
- HostingIdentity for hosted servers: extend your domain, build a new one, or use Entra IDWhich sign-in and directory design suits Windows servers moved to a hosting provider: extending the existing Active Directory domain, building a separate domain, or using Microsoft Entra services, with what each depends on.Read article
- TroubleshootingEntra Connect Sync Errors: Duplicate Attributes, Soft and Hard Match ConflictsClassify Entra Connect export errors (AttributeValueMustBeUnique, InvalidSoftMatch, InvalidHardMatch, LargeObject), fix the source object on-premises and verify on the next cycle.Read article
Risks and limits
- These commands have not been run or lab-tested.
- Event IDs, field names and values are from the Microsoft pages dated above and apply to the Windows versions they name.
- The pages omit the interval at which Entra Connect Sync runs; check your own schedule.
- The TPM rows on Microsoft's table that advise disabling the TPM are left out here because there is no safe back-out on the page.
- Federated environments add authentication-phase errors that are not covered here.
Sources
- Pending devices in Microsoft Entra ID — Microsoft Learn, page dated 2026-04-23, accessed 20 November 2026
- Troubleshoot Microsoft Entra hybrid joined devices — Microsoft Learn, page dated 2026-09-01, accessed 20 November 2026
- Troubleshoot devices by using the dsregcmd command — Microsoft Learn, page dated 2025-06-27, accessed 20 November 2026
- Configure Microsoft Entra hybrid join — Microsoft Learn, page dated 2025-06-27, accessed 20 November 2026
This article was originally published on .