Examkin

Question

Difficulty: MediumContainer-Based Workloads

A manufacturing company is designing a container-based quality control system on Azure. The system will run image classification models on images uploaded by factory cameras. The workloads have the following requirements:
- Scale up rapidly in response to a sudden influx of images uploaded to an Azure Blob Storage container.
- Scale down to zero when there are no images to process to minimize idle compute costs.
- Support event-driven scaling using KEDA without requiring the team to manage Kubernetes APIs, node pools, or cluster upgrades.

Which compute solution should you recommend?

  1. Azure Container AppsAnswer
  2. B
    Azure Kubernetes Service (AKS)
  3. C
    Azure Container Instances
  4. D
    Azure App Service (Linux)

Answer

Azure Container Apps
Azure Container Apps is the recommended solution because it is a serverless container hosting service designed for microservices and event-driven applications. It supports native autoscaling using KEDA, allowing containers to scale to zero when idle, and manages all underlying Kubernetes infrastructure, APIs, and upgrades, thereby minimizing administrative effort.

Step-by-Step Solution

1
Analyze the scaling and infrastructure management requirements.
The solution requires rapid scaling based on events (KEDA), scaling down to zero, and minimized administrative overhead.
This rules out solutions that require managing cluster infrastructure or lack native event-driven scaling mechanisms.
2
Compare candidate Azure container hosting services.
Azure Container Apps is fully managed, serverless, natively supports KEDA, and scales to zero. Azure Kubernetes Service (AKS) requires node and cluster administration. Azure Container Instances (ACI) lacks native KEDA autoscaling.
Choosing the service that balances KEDA scaling with low management overhead is key to meeting the design requirements.

Key Concept

Azure Container Apps provides a serverless platform for containerized applications, enabling KEDA-driven scaling and eliminating Kubernetes cluster management overhead.
Rate this question

Topics

Container-Based Workloads

More questions from this topic