A containerized ML model must scale GPU-backed inference on demand. Which services?
A. Container Apps + GPU nodes
B. AKS with GPU VM pool
C. ACI + GPU
D. App Service + GPU

For multi-tenant SaaS with tenant-level isolation (custom domains, certificates), which option is most suitable?
A. One Container App per tenant
B. AKS with namespaces
C. ACI per tenant
D. Service Fabric

You plan to migrate from on‑prem AKS to PaaS-managed container service with familiar Kubernetes underlying. Choose:
A. ACI
B. AKS
C. Container Apps
D. App Service

CI/CD zero-downtime deployment for Revision A→B in Container Apps uses:
A. Traffic splitting
B. Slot swap
C. Blue environment
D. Create new app

Choose the right service for container apps that require full orchestration and custom microservice policy.
A. Container Apps
B. AKS
C. ACI
D. Service Fabric