Microsoft Certified: Azure Administrator AssociateDeploy and manage Azure compute resourcesMedium
You manage an Azure Virtual Machine Scale Set (VMSS) that hosts a globally distributed, high-traffic web application. The application experiences fluctuating load throughout the day, with peak times requiring significantly more compute resources. You need to configure the VMSS to automatically scale out and scale in based on CPU utilization to optimize cost and performance.
- AManually adjust the instance count of the VMSS during peak and off-peak hours.
- BImplement a scheduled auto-scale rule that increases instances at fixed times and decreases them at other fixed times.
- CConfigure an auto-scale setting that scales out when average CPU utilization is above 70% and scales in when below 30%.
- DUse a custom metric from Azure Application Insights to trigger scaling actions.
Show answer & explanationAnswer & explanation
Correct answer: C. Configure an auto-scale setting that scales out when average CPU utilization is above 70% and scales in when below 30%.
Configuring auto-scale based on CPU utilization is a standard and effective way to dynamically adjust the number of VMSS instances to match the actual workload. Scaling out at 70% and in at 30% provides a good balance for responsiveness and cost optimization.
Why the other options are wrong
- A. Manual adjustment is inefficient and does not respond dynamically to load fluctuations.
- B. Scheduled scaling is less responsive to actual load changes and might over or under-provision resources.
- D. While possible, using a custom metric is more complex and not necessary when a standard metric like CPU utilization directly correlates with the stated problem of 'fluctuating load'.
Azure VMSS Auto-scale
A feature of Azure Virtual Machine Scale Sets that automatically adjusts the number of VM instances based on defined rules, such as CPU utilization, queue length, or schedules.
- Optimizes performance and cost by matching capacity to demand.
- Supports metric-based, schedule-based, and manual scaling.
- Can define both scale-out (increase instances) and scale-in (decrease instances) rules.
- Commonly uses CPU usage as a scaling metric.
Memory trick: Scale Sets Sense CPU Swings.