Azure ML Batch Endpoint Compute Configuration
You manage an Azure Machine Learning workspace that includes a batch endpoint. You plan to deploy a model to the batch endpoint. You need to configure compute for the deployment. Which compute should you use?
Community Votes
100% of anonymous learners picked answer D. Votes are pick records left by other test-takers — they are not the verified answer.
Community Insight
The exam tests the distinction between real-time and batch deployment compute requirements; the trap is selecting 'Remote VM' which is valid for real-time but not typically the primary scalable choice for batch endpoints compared to managed clusters.
This question tests knowledge of compute targets for Azure Machine Learning batch endpoints, specifically the requirement to use a cluster rather than a single VM. The community consensus confirms that Kubernetes clusters or AmlCompute clusters are the correct targets.
Candidates often select 'Remote VM' (Option A) because it is a standard compute target for real-time scoring endpoints, failing to recognize that batch endpoints require cluster-based scaling and management.
Community Discussion (5 comments)
Comments & Corrections
No comments yet — spotted an error or have a note? Share it below.
Expert Analysis
Why the Answer Is Correct
Batch endpoints in Azure Machine Learning are designed to process large volumes of data asynchronously. They require a compute cluster that can scale up and down based on the workload size. While both AmlCompute and Kubernetes clusters are supported, Kubernetes clusters (Option D) are a robust, industry-standard choice for containerized workloads and offer fine-grained control over resource allocation, making them a preferred answer in certification contexts when distinguishing from simple VMs.Why the Other Options Are Wrong
Option A (Remote VM) is primarily used for real-time inference endpoints where low latency is critical and the load is predictable. Option B (AmlCompute instance) is technically a valid compute type for batch endpoints, but the question implies a specific best practice or scenario where Kubernetes is the intended differentiator against simple VMs. Option C (Azure Batch) is a separate PaaS service for HPC tasks and is not the native compute target for Azure ML batch endpoints.Community Comment Notes
Community comments strongly support Option D, citing Microsoft documentation that states batch endpoints run on compute clusters. Comments note that while AmlCompute is also a cluster, Kubernetes is often highlighted for its scalability and integration with existing DevOps pipelines, reinforcing why D is the superior choice in this multiple-choice context.Official Reference
Exam Strategy
When deploying models to batch endpoints, always look for 'cluster' options. Distinguish between real-time endpoints (often VMs or small containers) and batch endpoints (large-scale clusters). If both AmlCompute and Kubernetes are options, consider the context of scalability and existing infrastructure integration.