↓ Skip to main content
  1. Home
  2. >
  3. Azure
  4. >
  5. AZ-104
  6. >
  7. This article

Azure AZ-104 Drill: VM Scale-Up Strategy - Managing Peak Compute Loads

·668 words·4 mins·
Architecture Drills Microsoft Azure AZ-104 Azure CAF Governance Virtual Machines
Jeff Taakey
Author
Jeff Taakey
21+ Year Enterprise Architect | Multi-Cloud Architect & Strategist.
Jeff's Architecture Insights
Go beyond static exam dumps. Jeff’s Insights is engineered to cultivate the mindset of a Production-Ready Architect. We move past ‘correct answers’ to dissect the strategic trade-offs and multi-cloud patterns required to balance reliability, security, and TCO in mission-critical environments.
Table of Contents

While preparing for the Azure AZ-104 exam, many candidates struggle with managing Azure Virtual Machine performance during predictable periodic workload spikes. In the enterprise world, this decision often hinges on balancing performance reliability vs. operational simplicity and cost. Let’s drill into a simulated migration scenario.

Scenario
#

Tailspin Manufacturing operates a line-of-business (LOB) application, AppLedger, hosted on a single Azure VM named VM-Prod. AppLedger handles critical financial reporting and does not support multiple active instances due to licensing constraints and data consistency requirements. At the end of every month, VM-Prod experiences significant CPU utilization spikes as the finance team runs intensive batch reports.

Key Requirements
#

Tailspin’s architecture and governance team needs to create an operational runbook that allows for temporarily increasing the compute capacity of VM-Prod during these monthly peak periods without impacting uptime or requiring app refactoring.

The Options
#

  • A) Install the Azure Monitor Performance Diagnostics extension on VM-Prod
  • B) Change the VM size of VM-Prod to a larger SKU during peak periods
  • C) Add VM-Prod to a Virtual Machine Scale Set (VMSS) for auto-scaling
  • D) Request an increase of the vCPU quota for the Azure subscription
  • E) Deploy Desired State Configuration (DSC) extension to VM-Prod

Correct Answer
#

B) Change the VM size of VM-Prod to a larger SKU during peak periods

The Architect’s Analysis
#

Correct Answer
#

Option B – resizing VM-Prod to a larger SKU for month-end batch processing.

Step-by-Step Winning Logic
#

AppLedger runs on a single instance VM that cannot be load-balanced or clustered due to application limitations. During monthly end-of-month batch jobs, CPU utilization spikes cause performance degradation. The optimal approach is to manually or programmatically resize the VM to a higher VM size SKU temporarily during the workload spike to meet performance demands. This approach aligns well with:

  • Performance Efficiency Pillar of the Well-Architected Framework: provision sufficient compute resources during peak loads.
  • Operational Excellence Pillar: Implement runbooks or Azure Automation to resize VM safely and avoid manual errors.
  • Cost Optimization Pillar: Avoid paying for higher VM SKU 100% of the time; pay more only when required.

The Traps (Distractor Analysis)
#

  • Option A: Installing performance diagnostic agents helps monitor but does not improve VM capacity.
  • Option C: Adding the VM to a scale set contradicts the application’s limitation—it does not support multiple active instances.
  • Option D: Increasing vCPU quota is necessary only if hitting subscription limits, but it does not directly improve VM-Prod performance temporarily.
  • Option E: DSC extension helps with configuration management but does not enable dynamic scaling or performance increase.

The Architect Blueprint
#

Diagram illustrating the VM resizing process triggered by operational runbook at month-end to upscale and then downscale after workloads normalize.

graph TD User([Finance Team]) --> |Runs Month-end Reports| VM-Prod RunbookRunner([Automation Runbook]) --> |Resize VM to Larger SKU| VM-Prod VM-Prod --> |Improved Compute Capacity| AppLedger RunbookRunner --> |Resize VM Back to Baseline SKU| VM-Prod style RunbookRunner fill:#0078D4,stroke:#333,color:#fff style VM-Prod fill:#5C2D91,stroke:#333,color:#fff

Diagram Note: The Automation Runbook orchestrates resizing the VM on-demand to meet performance needs during peak processing windows.

The Decision Matrix
#

Option Est. Complexity Est. Monthly Cost Impact Pros Cons
A) Install Performance Diagnostics Low Minimal Enables detailed monitoring No performance impact
B) Resize VM SKU Temporarily Medium Moderate during peak only Meets performance needs; cost efficient with runbook automation Requires downtime risk if not orchestrated properly
C) Add VM to Scale Set High High Supports auto-scaling Not supported by single-instance app, adds complexity
D) Increase vCPU Quota Low N/A (quota only) Removes subscription limits Does not solve immediate CPU bottleneck
E) DSC Extension Deployment Medium Minimal Helps with configuration drift Does not address performance scaling

Real-World Practitioner Insight
#

Exam Rule
#

“For the AZ-104 exam, select ‘Resize VM to Larger SKU’ when the app runs on a single VM instance and requires temporary compute scaling.”

Real World
#

“In enterprises, monthly batch jobs often necessitate scheduled VM resizing or burst scaling strategies. Azure Automation Runbooks paired with tags and schedules help automate resizing, minimizing manual effort and governance violations.”

Related Articles