AZ-104 Question 167
Single answerYour organization is launching a new containerized web application that experiences unpredictable spikes in traffic. You want a fully managed solution that automatically scales based on CPU usage and incoming requests, without having to manage underlying VMs or Kubernetes clusters. Which deployment approach best meets these requirements?
- A
Deploy using Azure Container Instances with manual CPU request changes
- B
Deploy to Azure Container Instances with autoscaling configured via Azure Monitor
- C
Deploy to Azure Container Apps with KEDA-based scale rules
- D
Deploy to an Azure Kubernetes Service cluster and configure a Horizontal Pod Autoscaler
Show answer and explanation
Correct answer: C
Explanation
Azure Container Apps provides automatic, event-driven scaling through KEDA without the overhead of managing a Kubernetes cluster. This makes it ideal for scenarios requiring rapid, demand-based scaling with minimal DevOps overhead.
- A. Incorrect.
Azure Container Instances with manual CPU changes does not provide automatic scaling; the admin must manually adjust resources.
- B. Incorrect.
Azure Monitor autoscaling for Container Instances is still not as seamless or built-in for on-demand scaling based on requests; it requires configuring alerts and custom logic.
- C. Correct.
Azure Container Apps support fully managed autoscaling via KEDA, which automatically scales containers based on CPU usage, HTTP requests, and other triggers without managing underlying infrastructure.
- D. Incorrect.
Azure Kubernetes Service with HPA offers powerful scaling but requires managing a Kubernetes cluster, which contradicts the requirement of minimal infrastructure management.