Microsoft AZ-400 Practice Test 95

Question 940

Another company wants warnings when CI failure rate or queue time exceeds a service delivery objective. Which unsupported assumption should the engineer challenge FIRST?

A query returning no rows is not proof there were no errors when its time filter excludes the incident.
Adding durations of parallel spans can overstate the actual end-to-end response time.
An error beginning near deployment does not alone prove that network routing caused it.
Production application uptime does not mean the software delivery system is healthy.

Explanation

Correct Answer

Production application uptime does not mean the software delivery system is healthy. This is the relevant unsupported assumption because the CI/CD platform is itself an operational service whose reliability deserves monitoring.

Incorrect Answers

This caution is valid in analyze metrics from instrumentation, but it is not the failure assumption illustrated by this scenario. The directly relevant risk is: Production application uptime does not mean the software delivery system is healthy.

Question 941

A microservice emits request counts but no distributed trace IDs, making it difficult to locate a failing downstream component. Which design or operational change MOST directly satisfies the requirement?

Propagate supported trace context and configure Application Insights distributed tracing across participating services.
Inspect trace parent-child relationships and the critical path rather than simply summing all spans.
Use consistent numerator and denominator definitions over the same time window and request population.
Use KQL summarize with a validated environment or run identifier and time filter at the correct event grain.

Explanation

Correct Answer

Distributed tracing enables causal investigation across service boundaries when context propagation is intact. The most directly responsive design or operational step is: Propagate supported trace context and configure Application Insights distributed tracing across participating services. Confirm the result by: Trace a test request and confirm service spans share the expected correlation and timing information.

Incorrect Answers

Distributed trace timing reflects concurrency and call structure; child durations may overlap. The relevant evidence instead is: Propagate supported trace context and configure Application Insights distributed tracing across participating services.

Service reliability ratios require clearly scoped compatible populations and time intervals. The acceptance criterion here is: Propagate supported trace context and configure Application Insights distributed tracing across participating services.

Log aggregation requires correct grouping keys to avoid duplicate or omitted incidents. What must be demonstrated is: Propagate supported trace context and configure Application Insights distributed tracing across participating services.

Question 942

After the team applies this response: Propagate supported trace context and configure Application Insights distributed tracing across participating services. In the situation—A microservice emits request counts but no distributed trace IDs, making it difficult to locate a failing downstream component. Which evidence BEST demonstrates the response is effective?

Compare raw event counts, alert exclusions and customer-impact incidents before and after the change.
Compare summarized failure counts with raw distinct run IDs and deployment audit records.
Trace a test request and confirm service spans share the expected correlation and timing information.
Compare raw run phase timestamps to the dashboard and identify the true critical-path stage.

Explanation

Correct Answer

The governing engineering principle is: Distributed tracing enables causal investigation across service boundaries when context propagation is intact.

Incorrect Answers

Reliable operational analysis requires visibility into exclusions that can change apparent performance. Instead, the response must trace a test request and confirm service spans share the expected correlation and timing information.

Log aggregation requires correct grouping keys to avoid duplicate or omitted incidents. The relevant control is to trace a test request and confirm service spans share the expected correlation and timing information.

Operational metrics should distinguish waiting for capacity from work actively being performed. The needed step here is to trace a test request and confirm service spans share the expected correlation and timing information.

Question 943

Another application chain spans API gateway, queue consumers and databases without end-to-end request correlation. Which engineering principle BEST explains the relevant requirement?

Operational metrics should distinguish waiting for capacity from work actively being performed.
Tail latency can reveal degraded user experience hidden by a healthy average.
Distributed tracing enables causal investigation across service boundaries when context propagation is intact.
Causal incident analysis needs corroborated telemetry rather than an inference from temporal coincidence.

Explanation

Correct Answer

The second engineering situation illustrates the control boundary captured by this principle: Distributed tracing enables causal investigation across service boundaries when context propagation is intact. The scenario is not resolved by an unrelated pipeline or policy activity.

Incorrect Answers

Operational metrics should distinguish waiting for capacity from work actively being performed. That principle concerns analyze metrics from instrumentation, whereas this specific incident is governed by: Distributed tracing enables causal investigation across service boundaries when context propagation is intact.

Tail latency can reveal degraded user experience hidden by a healthy average. That principle concerns analyze metrics from instrumentation, whereas this specific incident is governed by: Distributed tracing enables causal investigation across service boundaries when context propagation is intact.

The second engineering situation illustrates the control boundary captured by this principle: Causal incident analysis needs corroborated telemetry rather than an inference from temporal coincidence. The appropriate comparison is: Distributed tracing enables causal investigation across service boundaries when context propagation is intact.

Question 944

Another application chain spans API gateway, queue consumers and databases without end-to-end request correlation. Which unsupported assumption should the engineer challenge FIRST?

Fewer visible alerts do not necessarily mean the underlying application error rate has decreased.
A query returning no rows is not proof there were no errors when its time filter excludes the incident.
An error beginning near deployment does not alone prove that network routing caused it.
Independent log entries with similar timestamps do not automatically describe the same request.

Explanation

Correct Answer

Independent log entries with similar timestamps do not automatically describe the same request. This is the relevant unsupported assumption because distributed tracing enables causal investigation across service boundaries when context propagation is intact.

Incorrect Answers

This caution is valid in analyze metrics from instrumentation, but it is not the failure assumption illustrated by this scenario. The directly relevant risk is: Independent log entries with similar timestamps do not automatically describe the same request.

Question 945

A team configures an Azure Monitor alert rule but notification actions point to an empty action group. Which design or operational change MOST directly satisfies the requirement?

Use KQL summarize with a validated environment or run identifier and time filter at the correct event grain.
Separate queue latency, job execution and external dependency time when analyzing pipeline metrics.
Use a scoped KQL query that filters the relevant table by timestamp, service identifier and error conditions.
Bind alert rules to governed action groups with valid contacts or automation and test the delivery path.

Explanation

Correct Answer

An alert requires a working action path and responder to influence operational outcomes. The most directly responsive design or operational step is: Bind alert rules to governed action groups with valid contacts or automation and test the delivery path. Confirm the result by: Trigger an expected test alert and inspect fired status, action-group delivery and responsible acknowledgment.

Incorrect Answers

Log aggregation requires correct grouping keys to avoid duplicate or omitted incidents. The relevant evidence instead is: Bind alert rules to governed action groups with valid contacts or automation and test the delivery path.

Operational metrics should distinguish waiting for capacity from work actively being performed. The acceptance criterion here is: Bind alert rules to governed action groups with valid contacts or automation and test the delivery path.

KQL filtering is useful when query scope and timestamps match the intended evidence population. What must be demonstrated is: Bind alert rules to governed action groups with valid contacts or automation and test the delivery path.

Question 946

After the team applies this response: Bind alert rules to governed action groups with valid contacts or automation and test the delivery path. In the situation—A team configures an Azure Monitor alert rule but notification actions point to an empty action group. Which evidence BEST demonstrates the response is effective?

Trigger a relevant change and verify mandatory scan jobs execute and contribute to the status gate.
Trigger an expected test alert and inspect fired status, action-group delivery and responsible acknowledgment.
Trace a failed request to its authentication response, credential scope and release timestamp.
Compare raw event counts, alert exclusions and customer-impact incidents before and after the change.

Explanation

Correct Answer

The governing engineering principle is: An alert requires a working action path and responder to influence operational outcomes.

Incorrect Answers

Instrumentation must measure the presence of required checks as well as the outcome of executed checks. The appropriate comparison is: Trigger an expected test alert and inspect fired status, action-group delivery and responsible acknowledgment.

Causal incident analysis needs corroborated telemetry rather than an inference from temporal coincidence. The relevant evidence instead is: Trigger an expected test alert and inspect fired status, action-group delivery and responsible acknowledgment.

Reliable operational analysis requires visibility into exclusions that can change apparent performance. The acceptance criterion here is: Trigger an expected test alert and inspect fired status, action-group delivery and responsible acknowledgment.

Question 947

Another organization has CPU threshold alerts yet incidents continue because responders never receive notifications. Which engineering principle BEST explains the relevant requirement?

KQL filtering is useful when query scope and timestamps match the intended evidence population.
Log aggregation requires correct grouping keys to avoid duplicate or omitted incidents.
Resource saturation and application symptoms must be correlated before assigning root cause.
An alert requires a working action path and responder to influence operational outcomes.

Explanation

Correct Answer

The second engineering situation illustrates the control boundary captured by this principle: An alert requires a working action path and responder to influence operational outcomes. The scenario is not resolved by an unrelated pipeline or policy activity.

Incorrect Answers

KQL filtering is useful when query scope and timestamps match the intended evidence population. That principle concerns analyze metrics from instrumentation, whereas this specific incident is governed by: An alert requires a working action path and responder to influence operational outcomes.

Log aggregation requires correct grouping keys to avoid duplicate or omitted incidents. That principle concerns analyze metrics from instrumentation, whereas this specific incident is governed by: An alert requires a working action path and responder to influence operational outcomes.

Resource saturation and application symptoms must be correlated before assigning root cause. That principle concerns analyze metrics from instrumentation, whereas this specific incident is governed by: An alert requires a working action path and responder to influence operational outcomes.

Question 948

Another organization has CPU threshold alerts yet incidents continue because responders never receive notifications. Which unsupported assumption should the engineer challenge FIRST?

A green workflow summary is not proof every intended security check actually ran.
Dividing errors by only successful requests is not equivalent to errors divided by all attempted requests.
Counting every log message is not the same as counting distinct failed pipeline executions.
Creating a metric alert rule does not guarantee anyone is notified when it fires.

Explanation

Correct Answer

Creating a metric alert rule does not guarantee anyone is notified when it fires. This is the relevant unsupported assumption because an alert requires a working action path and responder to influence operational outcomes.

Incorrect Answers

This caution is valid in analyze metrics from instrumentation, but it is not the failure assumption illustrated by this scenario. The directly relevant risk is: Creating a metric alert rule does not guarantee anyone is notified when it fires.

Question 949

Azure Monitor ingestion includes production secrets because developers log full request headers for troubleshooting. Which design or operational change MOST directly satisfies the requirement?

Correlate container CPU, memory, throttling and application response metrics with scaling events.
Apply data minimization, redaction and controlled access to telemetry collection and storage.
Inspect required job execution coverage and distinguish success, failure, cancellation and skipped states.
Validate metric definitions, suppression policies and raw telemetry before treating reduced alerts as reliability improvement.

Explanation

Correct Answer

Instrumentation must protect sensitive data even when telemetry systems are used for legitimate diagnostics. The most directly responsive design or operational step is: Apply data minimization, redaction and controlled access to telemetry collection and storage. Confirm the result by: Generate a synthetic sensitive field and verify it is removed from collected logs and traces.

Incorrect Answers

Resource saturation and application symptoms must be correlated before assigning root cause. The appropriate remedy is to apply data minimization, redaction and controlled access to telemetry collection and storage.

Instrumentation must measure the presence of required checks as well as the outcome of executed checks. Instead, the response must apply data minimization, redaction and controlled access to telemetry collection and storage.

Reliable operational analysis requires visibility into exclusions that can change apparent performance. The relevant control is to apply data minimization, redaction and controlled access to telemetry collection and storage.

Question 950

After the team applies this response: Apply data minimization, redaction and controlled access to telemetry collection and storage. In the situation—Azure Monitor ingestion includes production secrets because developers log full request headers for troubleshooting. Which evidence BEST demonstrates the response is effective?

Trigger a relevant change and verify mandatory scan jobs execute and contribute to the status gate.
Generate a synthetic sensitive field and verify it is removed from collected logs and traces.
Compare summarized failure counts with raw distinct run IDs and deployment audit records.
Compare raw event counts, alert exclusions and customer-impact incidents before and after the change.

Explanation

Correct Answer

The governing engineering principle is: Instrumentation must protect sensitive data even when telemetry systems are used for legitimate diagnostics.

Incorrect Answers

Instrumentation must measure the presence of required checks as well as the outcome of executed checks. What must be demonstrated is: Generate a synthetic sensitive field and verify it is removed from collected logs and traces.

Log aggregation requires correct grouping keys to avoid duplicate or omitted incidents. The appropriate comparison is: Generate a synthetic sensitive field and verify it is removed from collected logs and traces.

Reliable operational analysis requires visibility into exclusions that can change apparent performance. The relevant evidence instead is: Generate a synthetic sensitive field and verify it is removed from collected logs and traces.

Designing and Implementing Microsoft DevOps Solutions (AZ-400) Practice Tests

Microsoft AZ-400 Practice Test 93 - 10 Questions (Last Reviewed 10-Oct-2026)
Microsoft AZ-400 Practice Test 94 - 10 Questions (Last Reviewed 09-Oct-2026)
Microsoft AZ-400 Practice Test 95 - 10 Questions (Last Reviewed 08-Oct-2026)
Microsoft AZ-400 Practice Test 96 - 10 Questions (Last Reviewed 07-Oct-2026)
Microsoft AZ-400 Practice Test 97 - 10 Questions (Last Reviewed 06-Oct-2026)

Pass Your Microsoft AZ-400 Exam Easy!

Question 940

Explanation

Correct Answer

Incorrect Answers

Question 941

Explanation

Correct Answer

Incorrect Answers

Question 942

Explanation

Correct Answer

Incorrect Answers

Question 943

Explanation

Correct Answer

Incorrect Answers

Question 944

Explanation

Correct Answer

Incorrect Answers

Question 945

Explanation

Correct Answer

Incorrect Answers

Question 946

Explanation

Correct Answer

Incorrect Answers

Question 947

Explanation

Correct Answer

Incorrect Answers

Question 948

Explanation

Correct Answer

Incorrect Answers

Question 949

Explanation

Correct Answer

Incorrect Answers

Question 950

Explanation

Correct Answer

Incorrect Answers

Designing and Implementing Microsoft DevOps Solutions (AZ-400) Practice Tests