AZ-700 Question 193
Single answerYou manage a critical internal application deployed on Azure Virtual Machines behind an Azure Load Balancer. The application experiences highly variable traffic spikes throughout the day. Management wants to ensure that the system can handle sudden surges without excessive overprovisioning during off-peak times. Which scaling approach would best address these requirements?
- A
Manually adjust the number of running VM instances at the start and end of peak hours
- B
Enable autoscale based on performance metrics, such as CPU usage or network throughput
- C
Use an Availability Set to ensure Azure automatically adds or removes VMs during peak hours
- D
Create a second deployment in another region to handle traffic overflow
Show answer and explanation
Correct answer: B
Explanation
Automated scaling (autoscale) is the appropriate choice for workloads with unpredictable or spiky traffic. Azure Monitor Autoscale can be configured to use performance metrics (like CPU usage or network traffic) to add or remove VM instances automatically, balancing costs against performance. Manual scaling works best for highly predictable workloads and requires constant human intervention. Availability Sets provide resiliency against planned and unplanned outages but do not scale resources. Deploying to another region can boost geo-redundancy but does not address dynamic scaling needs without an autoscaling configuration.
- A. Incorrect.
Option 1 is incorrect. Manual scaling requires constant monitoring and intervention. While it can work for predictable workloads, it is not efficient for highly variable demand because you risk either overprovisioning or failing to react quickly to spikes.
- B. Correct.
Option 2 is correct. Autoscale rules that track relevant performance metrics (e.g., CPU usage, memory, or network throughput) can dynamically increase or decrease instances in response to unpredictable traffic spikes. This method reduces manual overhead while ensuring availability.
- C. Incorrect.
Option 3 is incorrect. Availability Sets improve VM availability by distributing them across fault domains and update domains but do not automatically scale up or scale down the number of instances. They only help maintain uptime, not scaling.
- D. Incorrect.
Option 4 is incorrect. Deploying to another region can improve redundancy or geo-distribution, but it does not inherently provide automatic scaling in response to traffic spikes. You would still need an autoscaling mechanism.