View Full Microsoft AI-200 Exam Dumps and Practice Test Dumps

 

Question 361

Which Azure AI Search capability returns only selected document fields?

1. Search mode
2. Select clause
3. Query type
4. Scoring function

Correct Answer: 2

Explanation:

The $select parameter in Azure AI Search allows an application to specify which fields should be included in returned search documents. This can reduce unnecessary response data and ensure clients receive only the properties needed by the application. Search mode influences how search terms are interpreted, query type controls the query syntax or behavior, and scoring functions influence relevance calculations. Selecting only required fields can be useful in AI applications where search results are passed into downstream processing and unnecessary document content would increase response size.

Question 362

Which Cosmos DB feature provides automatic conflict resolution for multi-region writes?

1. Conflict resolution policy
2. Request throttling
3. Query metrics
4. Index transformation

Correct Answer: 1

Explanation:

Azure Cosmos DB conflict resolution policies determine how conflicting writes are handled in multi-region write configurations. Depending on the selected approach, conflicts can be resolved using supported automatic strategies or application-defined logic. Request throttling controls behavior when provisioned capacity is exceeded, query metrics provide information about query execution, and index transformation concerns indexing changes. Conflict resolution is important for globally distributed AI applications where several regional instances may update related data concurrently.

Question 363

Which Cosmos DB capability reports detailed request-unit consumption for queries?

1. Diagnostic logs
2. Query metrics
3. Change notifications
4. Partition statistics

Correct Answer: 2

Explanation:

Cosmos DB query metrics provide detailed information about query execution, including resource consumption characteristics that can help developers understand request-unit usage and query behavior. This information can assist with performance tuning and identifying expensive query patterns. Diagnostic logs provide broader operational information, change notifications expose modifications through supported mechanisms, and partition statistics provide partition-related information. Query metrics are particularly useful when optimizing AI applications that perform complex document retrieval and need to balance query performance against RU consumption.

Question 364

Which Blob Storage feature moves objects between access tiers automatically?

1. Lifecycle management
2. Version retention
3. Container recovery
4. Inventory policy

Correct Answer: 1

Explanation:

Azure Blob Storage lifecycle management can automatically move blobs between supported access tiers or delete them according to rule conditions such as age. This allows storage behavior to align with changing access patterns over an object’s lifetime. Version retention preserves earlier blob versions, container recovery protects against deletion, and inventory policies produce object reports. Lifecycle rules can be valuable for AI data platforms where recently generated datasets require frequent access but older training artifacts or historical files can transition to more economical storage.

Question 365

Which Azure Storage feature enables customer-controlled encryption keys?

1. Customer-managed key
2. Platform secret
3. Storage token
4. Service credential

Correct Answer: 1

Explanation:

Azure Storage supports customer-managed keys for scenarios where organizations require greater control over encryption-key management. Keys can be maintained through supported Azure key-management services, allowing customers to control key lifecycle operations such as rotation and access permissions. Platform-managed encryption remains available as the standard service-managed approach, while storage tokens and service credentials are authentication or authorization concepts rather than encryption-key controls. Customer-managed encryption can be important for AI workloads handling regulated or sensitive datasets with specific organizational security requirements.

Question 366

Which Service Bus feature moves repeatedly unsuccessful messages for separate inspection?

1. Dead-letter queue
2. Transfer queue
3. Session receiver
4. Message browser

Correct Answer: 1

Explanation:

The Service Bus dead-letter queue provides a separate location for messages that cannot be successfully processed or meet configured dead-letter conditions. Applications can inspect these messages later and determine whether they should be corrected, retried, or discarded. A transfer queue is associated with supported forwarding scenarios, a session receiver processes session-aware messages, and a message browser is not the standard Service Bus mechanism for isolating failed messages. Dead-lettering is important for resilient AI processing pipelines because problematic inputs can be separated from the normal processing stream.

Question 367

Which Service Bus capability keeps related messages associated with one processing context?

1. Session identifier
2. Broker sequence
3. Entity path
4. Correlation token

Correct Answer: 1

Explanation:

A Service Bus session identifier associates messages with a logical session so session-aware consumers can process related messages together. This can support ordered processing and stateful workflows where messages belong to the same business operation or conversation. Broker sequence numbers identify individual messages, entity paths identify Service Bus entities, and correlation tokens can be used for application-level relationship tracking. Session identifiers are useful for AI workflows involving ordered conversation events, job stages, or customer-specific processing sequences.

Question 368

Which Event Hubs setting identifies the logical destination within a namespace?

1. Event hub name
2. Namespace quota
3. Capture interval
4. Consumer offset

Correct Answer: 1

Explanation:

An Event Hub name identifies a specific event stream within an Event Hubs namespace. Applications publish and consume events through the named event hub, while the namespace provides the broader resource boundary containing one or more event hubs. Capture interval controls supported capture timing, consumer offsets represent processing positions, and namespace quotas relate to service capacity. Naming event hubs appropriately is useful when an AI platform separates streams by workload, such as application telemetry, model events, operational messages, or user activity.

Question 369

Which Event Grid concept connects a publisher to an event-processing destination?

1. Event subscription
2. Delivery endpoint
3. Publisher identity
4. Topic credential

Correct Answer: 1

Explanation:

An Event Grid event subscription defines how events from a topic or supported event source are delivered to a selected handler. It can include filtering and delivery-related configuration so that consumers receive appropriate events. A delivery endpoint represents where events are sent, while publisher identity and topic credential describe authentication or publishing concepts rather than the subscription object itself. Event subscriptions are fundamental to event-driven AI architectures because they connect resource or application events with processing functions, workflows, queues, or other supported handlers.

Question 370

Which Durable Functions pattern coordinates several independent activities before combining results?

1. Fan-out/fan-in
2. Chained execution
3. Human interaction
4. Monitor workflow

Correct Answer: 1

Explanation:

The Durable Functions fan-out/fan-in pattern starts multiple independent activities and later combines their results. It is useful when work can be performed concurrently but the workflow must wait for all required activities before proceeding. Chained execution handles sequential dependencies, human interaction patterns wait for external decisions, and monitor workflows repeatedly check conditions. Fan-out/fan-in is well suited to AI batch scenarios such as processing multiple documents, evaluating several inputs, or running parallel transformations before creating one consolidated result.

Question 371

Which Azure Functions trigger starts execution when a queue message arrives?

1. Queue trigger
2. Blob trigger
3. Timer trigger
4. Service endpoint

Correct Answer: 1

Explanation:

An Azure Functions queue trigger starts a function when a supported queue contains a message that meets the trigger’s processing conditions. This provides an event-driven mechanism for asynchronously processing queued workloads. A Blob trigger responds to supported blob changes, a Timer trigger invokes functions according to a schedule, and a service endpoint is a networking concept rather than a Functions trigger type. Queue-triggered functions are useful for AI workloads where incoming tasks need to be processed independently and asynchronously.

Question 372

Which Container Apps setting determines whether ingress is publicly reachable?

1. External ingress
2. Internal ingress
3. Private routing
4. Network exposure

Correct Answer: 2

Explanation:

Azure Container Apps supports internal ingress for applications that should be reachable only within the appropriate Container Apps environment and networking context rather than being directly exposed through public ingress. External ingress is designed for publicly reachable application endpoints, while private routing and network exposure are broader networking concepts. Internal ingress can be useful for AI microservices that communicate with one another while keeping backend inference, preprocessing, or orchestration components inaccessible from the public internet.

Question 373

Which ACR feature supports pulling images using an immutable content reference?

1. Digest pinning
2. Tag aliasing
3. Repository mirroring
4. Manifest labeling

Correct Answer: 1

Explanation:

Digest pinning means referencing a container image by its content digest rather than relying on a mutable tag. Because the digest corresponds to specific image content, it provides a stable reference for reproducible deployments. Tag aliasing provides human-readable references, repository mirroring concerns copying or synchronizing content, and manifest labeling describes metadata concepts. Digest-based deployment is valuable for AI services because a production inference environment can be configured to run the exact container artifact that was tested and approved.

Question 374

Which Azure Monitor destination stores collected logs for KQL analysis?

1. Log Analytics workspace
2. Metrics database
3. Dashboard repository
4. Alert archive

Correct Answer: 1

Explanation:

A Log Analytics workspace stores supported Azure Monitor log data and provides a centralized environment where that data can be queried using Kusto Query Language. It supports investigation, troubleshooting, analytics, and monitoring workflows across many Azure resources and applications. Metrics are handled through different monitoring mechanisms, dashboards visualize information, and alert archives are not the primary centralized query store for Azure Monitor logs. A workspace is especially useful for AI platforms that need to correlate application, infrastructure, and security telemetry.

Question 375

Which Azure API Management policy can add a value to an HTTP header?

1. Set-header
2. Forward-request
3. Cache-lookup
4. Rate-limit

Correct Answer: 1

Explanation:

The API Management set-header policy modifies HTTP request or response headers according to the configured policy expression. It can be used to add, replace, or otherwise control supported header values as traffic passes through the gateway. Forward-request controls backend request forwarding behavior, cache lookup policies handle caching operations, and rate-limit policies control request frequency. Header manipulation can support AI APIs that need to pass correlation identifiers, routing information, or application-specific metadata between gateway and backend services.

Question 376

Which Azure SQL capability distributes databases across shared compute resources?

1. Elastic pool
2. Hyperscale storage
3. Serverless database
4. Managed instance group

Correct Answer: 1

Explanation:

Azure SQL elastic pools allow multiple databases to share a common pool of compute resources. This model can be useful when individual databases have varying usage patterns and do not all require their maximum capacity simultaneously. Hyperscale is a database architecture designed for large-scale workloads, serverless adjusts compute for supported workload patterns, and managed instance groups are not the relevant Azure SQL Database pooling mechanism. Elastic pools can help applications containing many AI-related databases manage fluctuating workloads more efficiently.

Question 377

Which PostgreSQL setting controls the maximum number of client connections?

1. Connection limit
2. Backup retention
3. Replica count
4. Storage threshold

Correct Answer: 1

Explanation:

A PostgreSQL connection limit controls how many client connections can be established according to the applicable server configuration and service constraints. Managing connection counts is important because excessive concurrent connections can consume resources and negatively affect database performance. Backup retention controls how long supported backups are retained, replica count concerns replicated instances, and storage thresholds relate to storage capacity. AI applications that create many concurrent requests should manage connection behavior carefully, often using connection pooling where appropriate.

Question 378

Which Microsoft Entra capability evaluates contextual conditions before granting access?

1. Conditional Access
2. Application role
3. Group membership
4. Directory role

Correct Answer: 1

Explanation:

Microsoft Entra Conditional Access evaluates contextual signals and configured conditions before allowing access to protected applications or resources. Policies can consider supported factors such as user identity, device state, location, application, and authentication strength. Application roles define application-specific permissions, group membership associates identities with groups, and directory roles provide administrative permissions. Conditional Access can provide an additional security layer for AI applications by requiring stronger authentication or imposing access restrictions when specific conditions are met.

Question 379

Which Azure Policy component determines what happens when a rule matches?

1. Policy effect
2. Policy scope
3. Policy metadata
4. Policy parameter

Correct Answer: 1

Explanation:

The Azure Policy effect determines the behavior applied when a policy rule evaluates a resource or operation. Supported effects can include actions such as auditing, denying, modifying, or deploying required resources, depending on the policy definition. Policy scope identifies where the policy applies, metadata contains descriptive information, and parameters provide configurable values. Understanding policy effects is important when designing governance for AI environments because the same compliance requirement can be implemented as an observation-only control or as an enforcement mechanism.

Question 380

Which Bicep feature returns a value from a deployment?

1. Output declaration
2. Variable assignment
3. Resource loop
4. Parameter block

Correct Answer: 1

Explanation:

A Bicep output declaration returns a value from a deployment so that another deployment process, automation workflow, or dependent process can consume it. Outputs can expose values such as resource names, endpoints, or other calculated properties. Variables store reusable expressions inside the template, resource loops repeat declarations, and parameter blocks define deployment inputs. Outputs are useful in AI infrastructure deployments when newly created resources need to provide connection information or identifiers to subsequent automation stages without hard-coding those values.