Examkin

Question

Difficulty: HardScale Azure App Service Web Apps

A company hosts a financial processing API on an Azure App Service Web App that currently runs on the Standard (S1) pricing tier with a default instance count of 2. During peak hours, the API experiences high CPU utilization. You are configuring Azure Monitor autoscale rules to handle this load. You create a scale-out rule that increases the instance count by 1 when the average CPU Percentage is greater than 80% for 10 minutes. To prevent autoscale flapping (rapidly alternating between scaling out and scaling in), which of the following scale-in configurations should you implement?

  1. Decrease the instance count by 1 when the average CPU Percentage is less than 40% for 10 minutes.Answer
  2. B
    Decrease the instance count by 1 when the average CPU Percentage is less than 60% for 10 minutes.
  3. C
    Decrease the instance count by 1 when the average CPU Percentage is less than 75% for 10 minutes.
  4. D
    Scale down the App Service plan to the Free (F1) pricing tier when the average CPU Percentage is less than 40% for 10 minutes.

Answer

Decrease the instance count by 1 when the average CPU Percentage is less than 40% for 10 minutes.
The correct option correctly prevents flapping by establishing a scale-in threshold (40%) that is lower than the expected CPU per instance immediately after a scale-out event (approximately 54%). This ensures that the system will only scale back in when the overall workload significantly drops.

Step-by-Step Solution

1
Calculate the total CPU workload at the scale-out threshold.
Total workload = 2 instances×81% CPU=162%2 \text{ instances} \times 81\% \text{ CPU} = 162\% of a single instance's capacity.
To find the post-scale-out CPU utilization, we must first determine the total processing load that will be distributed among the new number of instances.
2
Calculate the expected CPU utilization per instance after scaling out.
Post-scale-out CPU = 162%/3 instances=54%162\% / 3 \text{ instances} = 54\% average CPU per instance.
When scaling out from 2 to 3 instances, the total load of 162% is divided among 3 instances.
3
Identify the threshold to avoid immediate scale-in.
The scale-in threshold must be strictly less than 54%54\%. A value of 40%40\% is appropriate.
If the scale-in threshold is higher than or equal to 54%54\% (such as 60%60\% or 75%75\%), the scale-in rule will trigger immediately after scaling out, causing flapping.

Key Concept

Autoscale flapping prevention in Azure App Service Plans
Estimated Time:2m 0s
Rate this question

Topics

Scale Azure App Service Web Apps

More questions from this topic