Domain 4.1: Design a compute solution

You are designing the network architecture for an Azure Kubernetes Service (AKS) cluster. The cluster will host hundreds of microservices. You have a strict requirement that every pod must receive an IP address from the Azure Virtual Network subnet so that on-premises resources connected via ExpressRoute can route traffic directly to individual pods. Which AKS network plugin must you use?

A development team uses Azure App Service to host a mission-critical web application. They want to implement a zero-downtime deployment strategy. When a new version of the application is released, they want to deploy it to a staging environment, run automated integration tests, and then seamlessly switch user traffic to the new version. If issues are detected, they must be able to instantly roll back. Which TWO features of Azure App Service should you use? (Select TWO)

You are designing a serverless architecture using Azure Functions. The function will process images uploaded to Blob Storage. The processing logic takes approximately 8 minutes to execute per image. The workload is highly variable, ranging from 0 to 1,000 images per hour. You need to ensure the function scales automatically based on demand, but you also need to avoid 'cold starts' because the application requires immediate processing when an image is uploaded. Which Azure Functions hosting plan should you recommend?

A company is deploying Azure Virtual Desktop (AVD) for 1,000 remote employees. To optimize costs, they are using multi-session Windows 11 host pools. Employees need their user profiles, application settings, and Outlook caches to persist across sessions, regardless of which session host VM they are routed to upon login. Which profile management solution should you implement?

A startup is building a microservices architecture using containers. They have 15 different microservices that need to communicate with each other. They want a fully managed serverless container service that supports Kubernetes Event-driven Autoscaling (KEDA), Distributed Application Runtime (Dapr), and scales to zero when there is no traffic. They do not want to manage underlying VMs or Kubernetes cluster upgrades. Which compute service should you recommend?

You are designing the network architecture for an Azure Kubernetes Service (AKS) cluster. Security requirements dictate that every pod must receive an IP address from the Azure Virtual Network subnet. You must also be able to apply Azure Network Security Groups (NSGs) and User Defined Routes (UDRs) directly to the pod IP addresses. Which AKS network plugin should you recommend?

Your development team is deploying a new version of a mission-critical web application hosted on Azure App Service. They require a deployment strategy that allows them to test the new version in the live production environment with 10% of real user traffic. If the new version performs well, they want to seamlessly switch 100% of traffic to it with zero downtime. Which Azure App Service feature should you use?

You are designing a serverless architecture using Azure Functions to process images uploaded to Blob Storage. The processing requires a custom Linux container image. The workload is highly unpredictable, with sudden massive spikes in uploads. The function must scale out rapidly to handle the spikes, but must not suffer from 'cold starts' (latency when a new instance spins up). Which Azure Functions hosting plan should you recommend?

You are designing an Azure Virtual Desktop (AVD) environment for 1,000 task workers. To minimize compute costs, you want multiple users to share the same virtual machine simultaneously. However, users must have a persistent profile experience; their settings and files must follow them regardless of which VM they log into. Which TWO components are required for this design? (Select TWO)

You are designing a compute architecture for a batch processing workload. The workload requires 50 virtual machines to process data from a queue. The processing is fault-tolerant; if a VM is terminated mid-process, the message returns to the queue and another VM picks it up. You need to design the most cost-effective compute solution possible. Which compute option should you recommend?