Endpoint Administrator MD-102 Practice Test 83
Question 820
An IT automation reads Intune managedDevice records by Microsoft Graph and must produce a report for one delegated operating region without inadvertently processing a similar-named device in another region. Which query-and-action design enforces the intended target boundary?
- Query stable managed-device IDs and validated approved group/region ownership before report processing, using the required API permissions and explicit scoped filters
- Match only the first six characters of device display names across all tenants
- Enable privileged remote Wipe for the reporting app to simplify list queries
- Read only locally cached results without checking record identity or staleness
Explanation
Correct Answer
A:Stable object identity and validated scope reduce accidental cross-region processing
Incorrect Answers
B:Display-name prefixes can collide and are not an authorization boundary
C:Destructive permissions do not improve the precision of read-only reporting
D:A stale, unverified cache can conflate former and current devices
Question 821
An IT automation reads Intune managedDevice records by Microsoft Graph and must produce a report for one delegated operating region without inadvertently processing a similar-named device in another region. A scheduled inventory job attributes two identically named devices to one office. What is the most plausible source of misattribution?
- The integration keyed records on mutable display name rather than unique managedDevice ID and verified regional attribution
- The output CSV includes a Unicode header
- A Defender antivirus engine updated overnight
- Both devices are Windows 11
Explanation
Correct Answer
A:Device names are nonunique and can be reused across different records
Incorrect Answers
B:Character encoding does not inherently merge two distinct device objects
C:Antivirus updates cannot duplicate object identity in the Graph results
D:Shared OS version does not by itself cause an identity collision
Question 822
An IT automation reads Intune managedDevice records by Microsoft Graph and must produce a report for one delegated operating region without inadvertently processing a similar-named device in another region. Which acceptance check confirms the reporting identity and region boundary?
- Run the report once against a single uniquely named device
- Process two distinct device IDs with the same display name in different approved-region cohorts and verify only the scoped ID appears in each output
- Compare only CSV row count to total licensed employees
- Confirm that the reporting application is visible in Enterprise applications
Explanation
Correct Answer
B:This tests stable identity and region filtering explicitly
Incorrect Answers
A:A one-device run cannot expose name collisions or cross-region leakage
C:Employee count is not per-device identity or region proof
D:App registration does not establish correct query filtering
Question 823
A PowerShell automation retrieves Intune managed-device inventory through Microsoft Graph. At peak time the API returns HTTP 429 with a Retry-After response header; the job must not flood the tenant. Which backoff strategy respects documented Graph throttling?
- Retry immediately 20 times in parallel to overcome throttling
- Discard the 429 and record the entire inventory as successfully refreshed
- Honor the Retry-After delay for failed requests and use bounded retries with paging and reduced polling as needed
- Elevate to tenant Global Administrator to bypass service quotas
Explanation
Correct Answer
C:Microsoft Graph specifies server-controlled retry timing and discourages immediate repeat calls
Incorrect Answers
A:Concurrent retries increase resource pressure while the throttle is active
B:A failed request cannot be claimed complete
D:Administrative role does not eliminate Microsoft Graph throttling
Question 824
A PowerShell automation retrieves Intune managed-device inventory through Microsoft Graph. At peak time the API returns HTTP 429 with a Retry-After response header; the job must not flood the tenant. A JSON batch call returns HTTP 200 overall yet two enclosed inventory requests return 429. What should the client do?
- Read only the outer Graph batch HTTP success status and skip per-operation response codes for the throttled actions
- Delete the entire Intune device inventory to decrease future request count
- Inspect each subrequest status, respect each failed request retry-after and replay only the throttled operations
- Omit all failed pages and continue reporting a complete dataset
Explanation
Correct Answer
C:Batched Graph requests are individually throttled despite an overall 200 envelope
Incorrect Answers
A:The batch envelope can succeed while individual operations fail or return 429; inspect subresponses and retry only throttled requests after appropriate delay.
B:Destructive deletion is unrelated to throttled read retries
D:Partial reads cannot be labeled complete inventory
Question 825
A PowerShell automation retrieves Intune managed-device inventory through Microsoft Graph. At peak time the API returns HTTP 429 with a Retry-After response header; the job must not flood the tenant. How should the integration be tested without driving additional rate-limit failures?
- Use a controlled 429 response fixture to verify bounded retry timing and final pagination completeness, including a batch with one failed subrequest
- Generate 500 simultaneous requests against production to confirm the service throttles
- Inspect only the application API permission consent page
- Manually repeat the full inventory while disregarding response headers
Explanation
Correct Answer
A:Deterministic throttling tests establish the algorithm without hammering production
Incorrect Answers
B:Deliberate production pressure is unnecessary and intrusive
C:Permissions do not prove retry behavior
D:Repeated unbounded requests are the failure pattern being prevented
Question 826
A desktop-support automation will run a protected remote action against Intune devices after a manager approves a scoped list. An agent recommendation is not itself an authorization. Which execution control keeps the action within the approved scope and permissions?
- Allow any read-only token to execute device Wipe when a report recommends it
- Require explicit authorized approval, exact managed-device ID matching and the specific privileged Graph remote-action permission before issuing a bounded action
- Use matching device display-name prefixes and skip confirmation for similarly named devices
- Let the automation retry destructive actions indefinitely until it sees a success message
Explanation
Correct Answer
B:Administrative approval, stable target identity and privileged authorization are separate requirements
Incorrect Answers
A:Read permission does not include destructive device actions
C:Names are not reliable stable identifiers for disruptive actions
D:Unbounded retries can duplicate device impact
Question 827
A desktop-support automation will run a protected remote action against Intune devices after a manager approves a scoped list. An agent recommendation is not itself an authorization. A Graph report suggests the approved device was retired, but a separate device with the same name was also affected. What control most likely failed?
- An Office app update channel was selected incorrectly
- The device compliance score changed between report runs
- A Windows app detection rule returned installed
- The automation did not bind the approved action to a unique managedDevice ID and validated permissioned target scope
Explanation
Correct Answer
D:Display-name-based targeting can affect the wrong device even if an approval exists
Incorrect Answers
A:Office servicing does not choose remote-action targets
B:Compliance measurement does not establish target identity
C:An application detection state is not action authorization
Question 828
A desktop-support automation will run a protected remote action against Intune devices after a manager approves a scoped list. An agent recommendation is not itself an authorization. Which preflight evidence should precede the actual action on one approved test device?
- A dashboard shows an anonymous device count without individual IDs
- The agent ranks an action as high confidence without an approval record
- Approval ties the exact managed-device ID, intended action, operator role, and allowed device scope, while an out-of-scope control ID is denied
- Approve a single managed-device action after checking only that the operator has a delegated Graph scope
Explanation
Correct Answer
C:This proves authorization and precise targeting before user-impacting changes
Incorrect Answers
A:Aggregate counts cannot prove safe action targeting
B:Prediction confidence is not human authorization
D:Delegated scope alone does not establish intended target device ID, resource scope and user/role authorization; require exact per-device approval and denial of an out-of-scope control.
Question 829
A manufacturer uses endpoint insights to recommend a mass restart following a crash spike; however the spike is limited to one driver-and-OS cohort and one safety-critical production shift. Which automation boundary should be required before taking the proposed action?
- Restart the entire Windows device cohort immediately after a reliability score worsens without correlating the implicated driver builds
- Suppress all app crash telemetry to reduce incident alerts
- Use driver and OS cohort evidence to scope a noncritical pilot, respect maintenance windows, and require explicit approval before mass action
- Roll back the same driver on every laptop before confirming whether the affected devices share the bad version
Explanation
Correct Answer
C:Endpoint insight is evidence for targeting, not authority to interrupt production systems
Incorrect Answers
A:A generic score is a triage signal, not proof that every device needs a disruptive restart; investigate affected driver/OS cohorts and a safe pilot.
B:Disabling evidence does not repair the cause
D:Driver rollback can be appropriate after diagnosis but indiscriminate rollback is unsafe before matching crash evidence to the actual driver/version cohort.
Question 830
A manufacturer uses endpoint insights to recommend a mass restart following a crash spike; however the spike is limited to one driver-and-OS cohort and one safety-critical production shift. Following a recommended restart, devices still crash and only the affected driver version reproduces the failure. What was the original inference error?
- The reliability service requires a desktop wallpaper refresh before evaluating crashes
- An operational symptom was mistaken for a root cause; restart recommendations did not address a driver-specific crash condition
- A restarted test device is briefly healthy, so presume the same reboot fixes the underlying driver defect for all clients
- A new Entra dynamic user group will eliminate the driver fault by itself
Explanation
Correct Answer
B:Rebooting a device is not equivalent to correcting the causative driver
Incorrect Answers
A:Wallpaper is not the failure mechanism
C:A temporary symptom improvement does not establish resolution of a driver-specific crash; inspect the driver event/version evidence and recurrence.
D:Group membership cannot fix a faulty installed driver