AI-200 — frequently asked questions

A Quick task is triggered immediately by events like Git commits or image pushes. A Schedule-triggered task runs at fixed intervals (e.g., daily) regardless of code changes.

Yes, but the question specifies minimizing external infrastructure and using ACR natively. ACR Tasks run entirely within Azure, whereas GitHub Actions rely on GitHub's runner infrastructure.

No, the Consumption plan only supports code-based deployments for Windows and Linux. Linux containers require the Premium plan.

The maximum execution duration can be configured up to 10 minutes. Beyond that, you need Premium or Dedicated plans.

Azure requires explicit registration for certain preview or specialized resource providers like Microsoft.App to ensure compliance and resource allocation before they can be used.

No, Azure Container Apps jobs must reside within a Managed Environment. You cannot create a job in a resource group directly without this intermediate infrastructure layer.

The request object is immutable after retrieval. You must explicitly call setBody to push your modified document back into the execution pipeline for storage.

Use the native JavaScript 'in' operator: if ('propertyName' in documentObject). This returns true if the property exists, regardless of its value.

Cosmos DB updates secondary indexes for every property change. More indexed properties mean more work per write operation.

No, as long as you include all properties used in WHERE clauses or JOINs, query performance remains optimal.

It records P95 latency and EXPLAIN ANALYZE for the same fixed workload, so later memory-pressure and cache-hit changes can be compared without peak-hour noise.

It can, because more memory per vCore enlarges the usable cache for the vector index; the identical P95/P99 benchmark proves whether disk reads actually fell.

Azure Functions provide a built-in Change Feed Processor that handles lease management and partition distribution automatically, whereas App Service requires manual implementation.

Push (Option A) uses triggers to invoke code automatically. Pull (Option D) uses FeedIterator to fetch changes programmatically, offering more control over batching and parallelism.

Setting the default TTL to -1 disables the automatic expiration feature for all items in the container, meaning items persist indefinitely until manually deleted.

Session Consistency guarantees that a read operation will always see the most recent committed version of an item written within the same session, even when reading from different regions.

Connection Policy configures static client behaviors like retry limits. Session tokens are dynamic values that change with every write and must be applied per-request via RequestOptions.

Different nodes will have different session views, potentially leading to stale reads or violations of monotonic read guarantees in a session-consistent account.

A Custom Topic allows you to publish your own events. A Partner Topic is pre-configured by Microsoft for specific SaaS providers (like GitHub) so you can subscribe to their events.

Yes, Domains allow multiple publishers to write to a single topic namespace, which might be better for multi-tenant scenarios, but Partner Topics remain for inbound SaaS events only.

Queue Storage is better for simple work items. Service Bus Topics/Subscriptions are required here because we need to filter messages for specific subsets of consumers (restaurants/drivers) from a single publisher.

Yes, if those drivers share the same filtering criteria (e.g., they both deliver for Restaurant A). However, if each driver has unique preferences, individual subscriptions or complex filtering rules are needed.

Key Vault is optional. App Service Application Settings are natively secure and separate from code, meeting the specific goal of avoiding Git exposure.

The container receives the setting as an environment variable at startup, allowing the application to read it without accessing the file system or repo.

Yes. Ingress must expose the app before Azure Container Apps can associate a custom hostname or certificate with it, so enable ingress first.

Add the custom domain name first so Azure generates the required verification values, publish those DNS records at your provider, then validate the domain and bind the certificate.

Session affinity is supported only when one revision handles all traffic; multiple revision mode can split requests across revisions, so the client-to-replica cookie mapping is not guaranteed.

No. TCP ingress operates at Layer 4 and does not inspect HTTP cookies, so it cannot implement the cookie-based session affinity that Container Apps provides through HTTP ingress.

It uses messageCount for the target queue depth; queueLength belongs to the Azure Storage Queue scaler, not the Service Bus scaler.

Keeping one replica warm avoids scale-to-zero cold starts, which the case requires when minimizing backend API cold-start delay.

DefaultAzureCredential includes fallbacks like environment variables and developer tools. If the exam requires 'no other authentication systems,' these fallbacks violate the constraint, making ManagedIdentityCredential the only compliant choice.

AddAzureAppConfiguration is an extension method that registers the Azure App Configuration settings provider with the .NET configuration system, allowing the app to load key-value pairs securely using the specified credential.

Encryption is applied at the store level using Customer-Managed Keys (CMK). When enabled, all values stored in Key-Value pairs are automatically encrypted.

No. While you can store layout configurations, Feature Flags are designed for dynamic toggling and segmentation. Key-Values require code changes to evaluate and apply variations.