Which Azure Compute Fits Each App Service Web App?

Why the Answer Is Correct

The item forces one recommendation per app from the drag-item list, and each app's description contains a signature phrase that maps cleanly to one Azure compute service. App1 is described by its trigger behaviour — work is handled when messages/events arrive — which is exactly the Azure Functions consumption model: event-driven, scale-to-zero, per-execution billing, no cluster to manage. App2's description centres on a multi-service, container-based application that needs orchestration, service discovery and independent scaling, which is the Azure Kubernetes Service (AKS) scenario rather than a PaaS web host. App3's description centres on large volumes of compute-intensive work that must be queued and distributed across many machines without a permanently running cluster, which is precisely what Azure Batch provides through pools, jobs and tasks. Matching trigger model (Functions), orchestration model (AKS) and job-scheduling model (Batch) is the grading logic, not the fact that all three are called 'web apps' in the stem.

Why Other Candidate Answers Are Wrong

The common wrong pairings come from ignoring the workload verb in each row. Assigning the event-driven app to AKS or Batch introduces a cluster or pool plus job plumbing for something that only needs a function triggered per message, which is over-engineered and costlier at low volume. Assigning the containerised multi-service app to Azure Functions forces you into a single deployment unit and loses per-service scaling, sidecars and standard Kubernetes tooling. Assigning the compute-intensive app to AKS means you own node pools, autoscaling and job cleanup yourself, while assigning it to Functions hits execution-timeout and per-instance limits for long-running parallel workloads. Also incorrect is reusing the same recommendation for every app — the instructions allow reuse only where the evidence supports it, and the three descriptions deliberately diverge. Plain App Service plans are a trap because the stem already says these are App Service web apps and asks for a compute recommendation, not a hosting-plan SKU.

Community Comment Notes

mesloth's concise mapping aligns exactly with this analysis, keying App1 to Azure Functions on the phrase "as they arrive" and then listing AKS for App2 and Azure Batch for App3. That keyword-driven reading is the right way to attack this item: the exhibits are dense, so scan for the trigger/scaling language rather than the app name. The comment is a single vote, so it is supporting evidence rather than proof — but it converges with the vendor's own service descriptions for event-driven, container-orchestrated and job-scheduled workloads.