Home › Certifications › Microsoft Azure Administrator AZ-104 › Compute — ARM/Bicep and Virtual Machines
43 multiple-choice questions and 24 flashcards on Compute — ARM/Bicep and Virtual Machines, about 15% of the Microsoft Azure Administrator AZ-104 bank. Every one carries a written rationale.
Compute — ARM/Bicep and Virtual Machines is one of 6 chapters in CoStudy's Microsoft Azure Administrator (AZ-104) bank, and it holds 43 of the bank's 280 multiple-choice questions — roughly 15% of the total. That proportion is not arbitrary: chapters follow the certifying body's published exam outline, and the number of questions in each is set by that domain's published weight, so the share of your practice time this chapter takes matches the share of the real exam it accounts for.
Studying by chapter is worth doing once you have a diagnostic score. A single overall percentage tells you whether you are close; it does not tell you which domain is dragging. Working a weak chapter in isolation, and re-testing it in isolation, is the fastest way to move a score that has stalled — and it is why the mock exams in CoStudy report by domain rather than as one number.
10 questions drawn from this chapter, with the full rationale shown — the controlling principle behind the right answer, and why each wrong option tempts and fails.
Before applying a change to a live resource group, an administrator wants a report of exactly which resources a template would create, modify or delete. Which operation should be run?
Answer: A — The what-if operation against the template with the intended deployment mode
A) Correct — what-if performs a preflight comparison of the template against current state and prints the predicted create, modify, delete and ignore results, and its output depends on the mode chosen. B) Deploying first and reviewing later is the opposite of a preview and risks deleting resources. C) Validate checks that the template is syntactically and structurally deployable; it does not enumerate property-level changes. D) Manual comparison is slow, error-prone and not a supported preview mechanism.
An autoscale rule on a scale set triggers scale-out repeatedly within a few minutes, overshooting the capacity actually needed. Which setting should be adjusted FIRST to stop the oscillation?
Answer: B — Lengthen the cool-down period so metrics stabilise before another action fires
B) Correct — the cool-down period suppresses further scaling actions while newly added instances start and begin absorbing load, which is the direct remedy for flapping. A) Raising the maximum lets the overshoot go further rather than preventing it. D) A shorter window makes the rule more reactive and would worsen the oscillation. C) Manual mode removes the symptom by removing autoscale altogether, which abandons the requirement.
Which VM size series is OPTIMIZED for memory-intensive workloads like in-memory databases?
Answer: C — E-series (memory-optimized)
E-series VMs are memory-optimized (high memory-to-vCPU ratio), suited for SAP HANA, large databases. F-series is compute-optimized (high CPU); D is balanced; B is burstable for low-utilization workloads; L is storage-optimized.
A VM is deployed without an availability set or zone. The single VM uses all Premium SSD disks. What's the SLA?
Answer: A — 99.9%
Single VM with Premium SSD or higher disks: 99.9%. Availability Set: 99.95%. Across 2+ AZs: 99.99%. Off-by-one SLA trap: 99.95% is for availability sets, not single VMs.
A VM is moved to a different region using Azure Resource Mover. What is NOT preserved automatically?
Answer: C — Public IP address.
Public IPs are regional; a new IP is assigned in the target region (or you bring/reassociate one). Off-by-one trap: candidates assume IP follows VM (it doesn't, it's a regional resource).
Which availability option places virtual machine instances in physically separate datacenters inside a single Azure region?
Answer: C — Availability zones, which place the instances in separate zones of the same region
C is the only construct that guarantees physical separation: each zone is an independent set of datacenters with its own power, cooling and network, which is why spreading instances across zones raises the SLA above the availability set figure. A is the classic trap, since fault and update domains separate racks and patch waves inside one datacenter rather than across datacenters. D deliberately does the opposite to cut latency. B has no redundancy, and E confines every instance to a single zone.
An Availability Set has 3 fault domains and 5 update domains. How many VMs minimum to maximize fault tolerance?
Answer: A — 3
VMs are distributed round-robin across fault domains first, then update domains. With 3 FDs and 2+ VMs, you get FD-redundancy. 3 VMs = each FD has at least one. Common misconception: need 5 (one per UD) or 15 (FD x UD).
A Spot VM is running. Azure needs the capacity back. What is the eviction notice period?
Answer: B — 30 seconds.
Spot VMs get a 30-second eviction notice via Scheduled Events. Apps must handle gracefully. Off-by-one cost trap: candidates expect 5 min (like AWS) but Azure is 30 seconds.
A governance team wants a reusable, versioned artefact stored in Azure that other teams can deploy without holding a copy of the template file. Which option is the MOST appropriate?
Answer: D — A template spec, which stores the template in a subscription as a versioned resource
D) Correct — a template spec is a first-class Azure resource holding a template and its versions, deployable by anyone with the right role and no local file. A) Azure Blueprints is deprecated in favour of template specs and deployment stacks, so it is not the supported answer. B) A parameter file supplies values but carries no template body. C) Exporting and distributing files is exactly the sharing problem the question asks to eliminate.
Which deployment option distributes VMs across hardware failure domains within a single datacenter?
Answer: A — Availability Sets
Availability Sets distribute VMs across fault domains (separate hardware/power) and update domains within ONE datacenter. Availability Zones distribute across PHYSICAL DATACENTERS within a region. VMSS adds auto-scaling; ASR is DR.
3 cards from the 24 in this chapter.
VM size families?
B (burstable), D (general), E (memory), F (compute), N (GPU), L (storage), H (HPC), M (memory).
Name the top-level elements of an ARM template.
$schema, contentVersion, apiProfile, parameters, variables, functions, resources and outputs. Only $schema, contentVersion and resources are required, though resources may be an empty collection.
Auto-shutdown VMs?
Set schedule via VM settings or DevTest Labs. Cost saving.
These are a sample. The full Compute — ARM/Bicep and Virtual Machines chapter runs 67 items with per-chapter progress tracking, on the web and in the iOS app.
Open Microsoft Azure Administrator AZ-104 in CoStudy →
All Microsoft Azure Administrator AZ-104 practice questions →