Question 830
An architect proposes this change: Use deployment slots with intentional sticky slot settings, smoke testing and a tested swap-back plan. The workload context is: An App Service team must release a new version while retaining an immediate rollback option and preventing staging secrets from swapping accidentally. Which evidence would most directly validate the change?
- Run an ARM/Bicep what-if and deployment against staging; compare resulting resources with the approved source.
- Shift test traffic to the canary revision and confirm the configured percentage and rollback.
- Rotate a test secret in Key Vault and verify the app resolves the new value after supported refresh behavior.
- Deploy to staging, verify slot-specific settings remain attached and perform controlled swap and rollback.
Explanation
Correct Answer
D:This observation directly verifies the intended design: Deploy to staging, verify slot-specific settings remain attached and perform controlled swap and rollback. The architecture being tested is: Use deployment slots with intentional sticky slot settings, smoke testing and a tested swap-back plan. The original symptom was: A slot swap exposes the staging database connection string to production.
Incorrect Answers
A:The architecture being tested is: Use deployment slots with intentional sticky slot settings, smoke testing and a tested swap-back plan.
B:That observation pertains to a separate failure: Every user immediately receives an untested new container image. Its cause is: Deployment has no revision-level incremental traffic control. It does not confirm the result of this design. The direct proof would be: Deploy to staging, verify slot-specific settings remain attached and perform controlled swap and rollback.
C:The alternative test targets: Use Azure Key Vault references in App Service settings with a managed identity authorized to read the referenced secrets. The scenario's actual proposed remedy is: Use deployment slots with intentional sticky slot settings, smoke testing and a tested swap-back plan. Verify that one by: Deploy to staging, verify slot-specific settings remain attached and perform controlled swap and rollback.
Question 831
During architecture troubleshooting, a team reports: A slot swap exposes the staging database connection string to production. Context: An App Service team must release a new version while retaining an immediate rollback option and preventing staging secrets from swapping accidentally. Which underlying design mistake explains the observed behavior most precisely?
- Infrastructure-as-code and environment parameterization were not enforced.
- Secrets were embedded in deployed app settings without vault-backed lifecycle management.
- Deployment has no revision-level incremental traffic control.
- Slot-sticky configuration was not used for environment-specific settings.
Explanation
Correct Answer
D:The incident is explained by: Slot-sticky configuration was not used for environment-specific settings. The reported symptom is: A slot swap exposes the staging database connection string to production. The corresponding design fix is: Use deployment slots with intentional sticky slot settings, smoke testing and a tested swap-back plan.
Incorrect Answers
A:The alternative diagnosis “Infrastructure-as-code and environment parameterization were not enforced” does not match the observed symptom: A slot swap exposes the staging database connection string to production. Context: An App Service team must release a new version while retaining an immediate rollback option and preventing staging secrets from swapping accidentally. The incident is explained by: Slot-sticky configuration was not used for environment-specific settings. The reported symptom is: A slot swap exposes the staging database connection string to production. The evidence instead supports “Slot-sticky configuration was not used for environment-specific settings”.
B:The alternative diagnosis “Secrets were embedded in deployed app settings without vault-backed lifecycle management” does not match the observed symptom: A slot swap exposes the staging database connection string to production. Context: An App Service team must release a new version while retaining an immediate rollback option and preventing staging secrets from swapping accidentally. The incident is explained by: Slot-sticky configuration was not used for environment-specific settings. The reported symptom is: A slot swap exposes the staging database connection string to production. The evidence instead supports “Slot-sticky configuration was not used for environment-specific settings”.
C:The alternative diagnosis “Deployment has no revision-level incremental traffic control” does not match the observed symptom: A slot swap exposes the staging database connection string to production. Context: An App Service team must release a new version while retaining an immediate rollback option and preventing staging secrets from swapping accidentally. The incident is explained by: Slot-sticky configuration was not used for environment-specific settings. The reported symptom is: A slot swap exposes the staging database connection string to production. The evidence instead supports “Slot-sticky configuration was not used for environment-specific settings”.
Question 832
The design change is: Use deployment slots with intentional sticky slot settings, smoke testing and a tested swap-back plan. The original failure was: A slot swap exposes the staging database connection string to production. Which ongoing control best prevents recurrence of the same root cause?
- Require template review and deployment validation before each release.
- Recheck slot setting flags and protected connection strings before release.
- Monitor Key Vault reference errors and managed identity access after app slot swaps.
- Monitor canary health and revision deactivation policy after full rollout.
Explanation
Correct Answer
B:This is an ongoing protection specific to the fault: Recheck slot setting flags and protected connection strings before release. The recurrence risk is: Slot-sticky configuration was not used for environment-specific settings. It should be checked using: Deploy to staging, verify slot-specific settings remain attached and perform controlled swap and rollback.
Incorrect Answers
A:The alternative diagnosis “Require template review and deployment validation before each release” does not match the observed symptom: The design change is: Use deployment slots with intentional sticky slot settings, smoke testing and a tested swap-back plan. The original failure was: A slot swap exposes the staging database connection string to production. Which ongoing control best prevents recurrence of the same root cause?. This is an ongoing protection specific to the fault: Recheck slot setting flags and protected connection strings before release. The recurrence risk is: Slot-sticky configuration was not used for environment-specific settings. The evidence instead supports “Recheck slot setting flags and protected connection strings before release”.
C:The alternative diagnosis “Monitor Key Vault reference errors and managed identity access after app slot swaps” does not match the observed symptom: The design change is: Use deployment slots with intentional sticky slot settings, smoke testing and a tested swap-back plan. The original failure was: A slot swap exposes the staging database connection string to production. Which ongoing control best prevents recurrence of the same root cause?. This is an ongoing protection specific to the fault: Recheck slot setting flags and protected connection strings before release. The recurrence risk is: Slot-sticky configuration was not used for environment-specific settings. The evidence instead supports “Recheck slot setting flags and protected connection strings before release”.
D:The alternative diagnosis “Monitor canary health and revision deactivation policy after full rollout” does not match the observed symptom: The design change is: Use deployment slots with intentional sticky slot settings, smoke testing and a tested swap-back plan. The original failure was: A slot swap exposes the staging database connection string to production. Which ongoing control best prevents recurrence of the same root cause?. This is an ongoing protection specific to the fault: Recheck slot setting flags and protected connection strings before release. The recurrence risk is: Slot-sticky configuration was not used for environment-specific settings. The evidence instead supports “Recheck slot setting flags and protected connection strings before release”.
Question 833
A design workshop has this requirement: A container service wants to direct 5% of live traffic to a new revision while keeping rollback available without changing DNS endpoints. Operators currently observe: Every user immediately receives an untested new container image. Select the ONE design remedy that best addresses the root cause.
- Use Azure Event Hubs with partitioned event streams, independent consumer groups and checkpointed readers.
- Use Azure Container Apps revision traffic splitting with an appropriate multiple-revision mode and routing.
- Use deployment slots with intentional sticky slot settings, smoke testing and a tested swap-back plan.
- Evaluate Azure Container Apps Dapr integration with compatible sidecar components for application APIs.
Explanation
Correct Answer
B:The targeted implementation is: Use Azure Container Apps revision traffic splitting with an appropriate multiple-revision mode and routing. The problem is not generic resource availability: Deployment has no revision-level incremental traffic control. Verify the result by: Shift test traffic to the canary revision and confirm the configured percentage and rollback.
Incorrect Answers
A:The alternative diagnosis “Use Azure Event Hubs with partitioned event streams, independent consumer groups and checkpointed readers” does not match the observed symptom: Every user immediately receives an untested new container image. The targeted implementation is: Use Azure Container Apps revision traffic splitting with an appropriate multiple-revision mode and routing. The problem is not generic resource availability: Deployment has no revision-level incremental traffic control. The evidence instead supports “Use Azure Container Apps revision traffic splitting with an appropriate multiple-revision mode and routing”.
C:The alternative diagnosis “Use deployment slots with intentional sticky slot settings, smoke testing and a tested swap-back plan” does not match the observed symptom: Every user immediately receives an untested new container image. The targeted implementation is: Use Azure Container Apps revision traffic splitting with an appropriate multiple-revision mode and routing. The problem is not generic resource availability: Deployment has no revision-level incremental traffic control. The evidence instead supports “Use Azure Container Apps revision traffic splitting with an appropriate multiple-revision mode and routing”.
D:The alternative diagnosis “Evaluate Azure Container Apps Dapr integration with compatible sidecar components for application APIs” does not match the observed symptom: Every user immediately receives an untested new container image. The targeted implementation is: Use Azure Container Apps revision traffic splitting with an appropriate multiple-revision mode and routing. The problem is not generic resource availability: Deployment has no revision-level incremental traffic control. The evidence instead supports “Use Azure Container Apps revision traffic splitting with an appropriate multiple-revision mode and routing”.
Question 834
The planned architecture is: Use Azure Container Apps revision traffic splitting with an appropriate multiple-revision mode and routing. Workload context: A container service wants to direct 5% of live traffic to a new revision while keeping rollback available without changing DNS endpoints. Which acceptance test demonstrates that the required behavior works?
- Invoke a Dapr service and confirm scoped access to the configured state or pub/sub component.
- Shift test traffic to the canary revision and confirm the configured percentage and rollback.
- Run two consumer groups and verify each can independently replay and checkpoint a known event batch.
- Deploy to staging, verify slot-specific settings remain attached and perform controlled swap and rollback.
Explanation
Correct Answer
B:The decisive evidence is: Shift test traffic to the canary revision and confirm the configured percentage and rollback. This measures the mechanism of the planned change, not an unrelated Azure service: Use Azure Container Apps revision traffic splitting with an appropriate multiple-revision mode and routing.
Incorrect Answers
A:That observation pertains to a separate failure: Teams repeatedly hard-code broker URLs into every service and create configuration drift. Its cause is: Common sidecar messaging/state integration components were not used where supported. It does not confirm the result of this design. The direct proof would be: Shift test traffic to the canary revision and confirm the configured percentage and rollback.
C:This measures the mechanism of the planned change, not an unrelated Azure service: Use Azure Container Apps revision traffic splitting with an appropriate multiple-revision mode and routing.
D:That observation pertains to a separate failure: A slot swap exposes the staging database connection string to production. Its cause is: Slot-sticky configuration was not used for environment-specific settings. It does not confirm the result of this design. The direct proof would be: Shift test traffic to the canary revision and confirm the configured percentage and rollback.
Question 835
The customer states: A container service wants to direct 5% of live traffic to a new revision while keeping rollback available without changing DNS endpoints. An incident reveals: Every user immediately receives an untested new container image. What cause best accounts for this evidence?
- Deployment has no revision-level incremental traffic control.
- Common sidecar messaging/state integration components were not used where supported.
- A competing-consumer queue was used instead of an append-style partitioned event stream.
- Slot-sticky configuration was not used for environment-specific settings.
Explanation
Correct Answer
A:This is the failure mechanism rather than an incidental symptom: Deployment has no revision-level incremental traffic control. It accounts for: Every user immediately receives an untested new container image. The architectural repair is: Use Azure Container Apps revision traffic splitting with an appropriate multiple-revision mode and routing.
Incorrect Answers
B:Sidecar configuration would not address the need for deployment has no revision-level incremental traffic control.
C:It accounts for: Every user immediately receives an untested new container image.
D:Slot-sticky configuration was not used for environment-specific settings would not address the need for deployment has no revision-level incremental traffic control.
Question 836
Workload context: A container service wants to direct 5% of live traffic to a new revision while keeping rollback available without changing DNS endpoints. The implemented remedy is: Use Azure Container Apps revision traffic splitting with an appropriate multiple-revision mode and routing. Which ongoing practice most specifically prevents recurrence of this failure: Every user immediately receives an untested new container image.?
- Monitor partitions, throughput, consumer lag and checkpoint storage after scaling devices.
- Review Dapr component scopes, secret stores and service identity as applications evolve.
- Monitor canary health and revision deactivation policy after full rollout.
- Recheck slot setting flags and protected connection strings before release.
Explanation
Correct Answer
C:The preventive control is: Monitor canary health and revision deactivation policy after full rollout. It targets the earlier cause: Deployment has no revision-level incremental traffic control. Confirm ongoing effectiveness by: Shift test traffic to the canary revision and confirm the configured percentage and rollback.
Incorrect Answers
A:Monitoring partitions, throughput, consumer lag and checkpoint storage after scaling devices would not monitor canary health and revision deactivation policy after full rollout.
B:Reviewing Dapr component scopes, secret stores and service identity as applications evolve would not monitor canary health and revision deactivation policy after full rollout.
D:Recheck slot setting flags and protected connection strings would not monitor canary health and revision deactivation policy after full rollout.
Question 837
A solution architect is reviewing a workload: A business API needs to decouple interactive request acknowledgments from time-consuming asynchronous invoice processing. The observed incident is: Client HTTP requests frequently time out because processing runs synchronously for several minutes. Which architecture choice most directly removes the identified weakness?
- Use an API frontend to accept validated commands into a durable Service Bus queue, with independent scalable workers.
- Design Azure Front Door with origin groups, routes and WAF policy as appropriate.
- Use Azure Service Bus queue with appropriate lock, retry, dead-letter and idempotency design.
- Bind an Azure Functions Service Bus trigger with a compatible hosting plan and idempotent processing.
Explanation
Correct Answer
A:The targeted implementation is: Use an API frontend to accept validated commands into a durable Service Bus queue, with independent scalable workers. The problem is not generic resource availability: Application processing and frontend request lifetime are tightly coupled. Verify the result by: Post a test command, observe fast 202/acknowledgment and eventual idempotent worker completion.
Incorrect Answers
B:Designing Azure Front Door with origin groups, routes and WAF policy as appropriate would not meet the need for an API frontend to accept validated commands into a durable Service Bus queue, with independent scalable workers.
C:Using Azure Service Bus queue with appropriate lock, retry, dead-letter and idempotency design would not meet the need for an API frontend to accept validated commands into a durable Service Bus queue, with independent scalable workers.
D:Bind an Azure Functions Service Bus trigger with a compatible hosting plan and idempotent processing would not meet the need for an API frontend to accept validated commands into a durable Service Bus queue, with independent scalable workers.
Question 838
An architect proposes this change: Use an API frontend to accept validated commands into a durable Service Bus queue, with independent scalable workers. The workload context is: A business API needs to decouple interactive request acknowledgments from time-consuming asynchronous invoice processing. Which evidence would most directly validate the change?
- Post a test command, observe fast 202/acknowledgment and eventual idempotent worker completion.
- Send requests to defined paths and confirm origin selection and WAF blocking in logs.
- Send a test invoice command and confirm the function consumes and settles it according to retry settings.
- Publish a sample command, force one processing failure and verify controlled retry or dead-letter behavior.
Explanation
Correct Answer
A:The decisive evidence is: Post a test command, observe fast 202/acknowledgment and eventual idempotent worker completion. This measures the mechanism of the planned change, not an unrelated Azure service: Use an API frontend to accept validated commands into a durable Service Bus queue, with independent scalable workers.
Incorrect Answers
B:The decisive evidence is: Post a test command, observe fast 202/acknowledgment and eventual idempotent worker completion.
C:That observation pertains to a separate failure: A 24/7 virtual machine polls a queue inefficiently while the managed function is idle. Its cause is: The integration omitted an event-triggered compute consumer. It does not confirm the result of this design. The direct proof would be: Post a test command, observe fast 202/acknowledgment and eventual idempotent worker completion.
D:The alternative test targets: Use Azure Service Bus queue with appropriate lock, retry, dead-letter and idempotency design. The scenario's actual proposed remedy is: Use an API frontend to accept validated commands into a durable Service Bus queue, with independent scalable workers. Verify that one by: Post a test command, observe fast 202/acknowledgment and eventual idempotent worker completion.
Question 839
During architecture troubleshooting, a team reports: Client HTTP requests frequently time out because processing runs synchronously for several minutes. Context: A business API needs to decouple interactive request acknowledgments from time-consuming asynchronous invoice processing. Which underlying design mistake explains the observed behavior most precisely?
- Global application entry routing and edge filtering were not implemented.
- The integration omitted an event-triggered compute consumer.
- The integration lacks durable command messaging and failure handling.
- Application processing and frontend request lifetime are tightly coupled.
Explanation
Correct Answer
D:The recorded behavior follows from: Application processing and frontend request lifetime are tightly coupled. The appropriate mitigation would be: Use an API frontend to accept validated commands into a durable Service Bus queue, with independent scalable workers. Verify with: Post a test command, observe fast 202/acknowledgment and eventual idempotent worker completion.
Incorrect Answers
A:The alternative diagnosis “Global application entry routing and edge filtering were not implemented” does not match the observed symptom: Client HTTP requests frequently time out because processing runs synchronously for several minutes. Context: A business API needs to decouple interactive request acknowledgments from time-consuming asynchronous invoice processing. The recorded behavior follows from: Application processing and frontend request lifetime are tightly coupled. The appropriate mitigation would be: Use an API frontend to accept validated commands into a durable Service Bus queue, with independent scalable workers. The evidence instead supports “Application processing and frontend request lifetime are tightly coupled”.
B:The alternative diagnosis “The integration omitted an event-triggered compute consumer” does not match the observed symptom: Client HTTP requests frequently time out because processing runs synchronously for several minutes. Context: A business API needs to decouple interactive request acknowledgments from time-consuming asynchronous invoice processing. The recorded behavior follows from: Application processing and frontend request lifetime are tightly coupled. The appropriate mitigation would be: Use an API frontend to accept validated commands into a durable Service Bus queue, with independent scalable workers. The evidence instead supports “Application processing and frontend request lifetime are tightly coupled”.
C:The alternative diagnosis “The integration lacks durable command messaging and failure handling” does not match the observed symptom: Client HTTP requests frequently time out because processing runs synchronously for several minutes. Context: A business API needs to decouple interactive request acknowledgments from time-consuming asynchronous invoice processing. The recorded behavior follows from: Application processing and frontend request lifetime are tightly coupled. The appropriate mitigation would be: Use an API frontend to accept validated commands into a durable Service Bus queue, with independent scalable workers. The evidence instead supports “Application processing and frontend request lifetime are tightly coupled”.
Question 840
A design review identifies: A business API needs to decouple interactive request acknowledgments from time-consuming asynchronous invoice processing. The failure is: Client HTTP requests frequently time out because processing runs synchronously for several minutes. Select TWO distinct decisions that together provide the immediate technical solution AND an ongoing prevention control. A verification-only step does not replace either responsibility.
- Bind an Azure Functions Service Bus trigger with a compatible hosting plan and idempotent processing.
- Monitor dead-letter count and duplicate processing attempts under worker changes.
- Design Azure Front Door with origin groups, routes and WAF policy as appropriate.
- Monitor queue age and retry backlog separately from frontend latency.
- Use an API frontend to accept validated commands into a durable Service Bus queue, with independent scalable workers.
Explanation
Correct Answers
D:The key follow-up control is: Monitor queue age and retry backlog separately from frontend latency. It avoids recurrence of the scenario in which: Client HTTP requests frequently time out because processing runs synchronously for several minutes. It addresses: Application processing and frontend request lifetime are tightly coupled.
E:This option addresses the actual design defect: Application processing and frontend request lifetime are tightly coupled. The correction is: Use an API frontend to accept validated commands into a durable Service Bus queue, with independent scalable workers. Confirm the outcome with: Post a test command, observe fast 202/acknowledgment and eventual idempotent worker completion.
Incorrect Answers
A:A Service Bus-triggered Function can be a worker, but binding a trigger alone does not change the synchronous API into an enqueue-and-acknowledge workflow.
B:Dead-letter and duplicate monitoring are useful operations controls but do not independently split API acknowledgement from durable asynchronous invoice processing.
C:Front Door improves edge routing and protection; it cannot solve long-running synchronous invoice work inside the API request path.
Designing Microsoft Azure Infrastructure Solutions (AZ-305) Practice Tests