AZ-900 Describe Azure architecture and services Practice Question

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

• ✗

  Azure Elastic Pool

  Why it's wrong here

  Azure Elastic Pool is a feature specific to Azure SQL Database that allows multiple databases to share a fixed pool of eDTU/vCore resources, providing predictable performance and cost efficiency for databases with variable usage patterns. It does not apply to general Azure compute resources like virtual machines or App Services, and it does not dynamically add or remove servers or instances in response to load. While both Elastic Pools and Autoscale involve resource management, Elastic Pools only govern SQL database resource allocation within a predefined limit, not the automatic scaling of compute capacity. Hence, it is not the correct answer for automatically adjusting compute resources based on metrics.

• ✓

  Azure Autoscale

  Why this is correct

  Azure Autoscale is the correct answer because it is the built-in Azure service that automatically increases or decreases the number of instances of a resource—such as virtual machine scale sets, App Service plans, or Azure Functions—based on metric conditions (e.g., CPU percentage, queue length) or a fixed schedule. This ensures that applications handle varying workloads efficiently without manual intervention, scaling out under load and scaling back in during off-peak times to reduce cost. Autoscale rules define thresholds and cooldown periods to prevent rapid flapping, and it is managed through Azure Monitor or resource-specific settings. Therefore, it directly fulfills the question's description of automatically adjusting compute resources.

• ✗

  Azure Load Balancer

  Why it's wrong here

  Azure Load Balancer distributes incoming network traffic across multiple healthy instances, such as VMs or services, using layer-4 (TCP/UDP) rules, thereby improving availability and responsiveness of the application. However, it does not change the number of instances; it only spreads existing traffic among whatever compute resources are currently deployed. Scaling the number of instances is a separate operational task that Autoscale performs based on load metrics. Thus, Load Balancer is incorrect because it focuses on traffic distribution rather than on dynamically increasing or decreasing the resource pool capacity.

• ✗

  Azure Traffic Manager

  Why it's wrong here

  Azure Traffic Manager operates at the DNS level to route end-user traffic to different service endpoints based on routing methods such as priority, performance, or geographic location. It does not modify the number of running compute instances; it simply directs incoming requests to healthy, pre-existing endpoints. Autoscale, in contrast, actively adjusts the instance count of a resource (such as a VM scale set or App Service plan) based on metric thresholds or a defined schedule. Therefore, Traffic Manager is incorrect because it manages traffic routing rather than resource quantity.

Go deeper

Related to this question