Examkin

Question

Difficulty: HardScale Azure App Service Web Apps

You manage a web application named ShipRoute that is currently hosted on a Basic (B1B1) App Service plan. During seasonal promotions, the application experiences high CPU utilization. You must configure autoscale rules to automatically increase the instance count by 11 when the average CPU Percentage is greater than 75%75\% for 1010 minutes. You also need to scale in the application by decreasing the instance count by 11 when CPU utilization drops, while ensuring that the scale-in action does not immediately trigger flapping when the instance count scales out from 11 to 22.

What configuration should you apply?

  1. Scale up the App Service plan to the Standard (S1S1) tier, and configure a scale-in rule with a CPU Percentage threshold of less than 35%35\%.Answer
  2. B
    Scale up the App Service plan to the Standard (S1S1) tier, and configure a scale-in rule with a CPU Percentage threshold of less than 50%50\%.
  3. C
    Keep the App Service plan on the Basic (B1B1) tier, and configure a scale-in rule with a CPU Percentage threshold of less than 35%35\%.
  4. D
    Keep the App Service plan on the Basic (B1B1) tier, and configure a scale-in rule with a CPU Percentage threshold of less than 50%50\%.

Answer

Scale up the App Service plan to the Standard (S1S1) tier, and configure a scale-in rule with a CPU Percentage threshold of less than 35%35\%.
Upgrading the App Service plan to the Standard (S1S1) tier is required because the Basic (B1B1) tier does not support custom autoscale rules. To prevent flapping when the instance count increases from 11 to 22 at a scale-out threshold of 75%75\% CPU, the scale-in CPU threshold must be set below the resulting average CPU load of 37.5%37.5\% (calculated as 75%/275\% / 2). Setting the scale-in rule to trigger when the CPU Percentage is less than 35%35\% ensures the rule is not immediately triggered after a scale-out event.

Step-by-Step Solution

1
Determine the minimum App Service plan tier required for autoscale rules.
Standard (S1S1) tier or higher.
The Basic (B1B1) tier only supports manual scaling up to 33 instances, whereas the Standard (S1S1) tier is the entry level for metric-based and schedule-based autoscale.
2
Calculate the average CPU Percentage per instance immediately after scaling out from 11 to 22 instances.
Approximately 37.5%37.5\%.
The scale-out rule triggers when the CPU exceeds 75%75\% on the single instance. Assuming the total load remains constant at 75%75\% during the transition, distributing this load across 22 instances yields an average of 75%/2=37.5%75\% / 2 = 37.5\% CPU per instance.
3
Set the scale-in threshold to a value that prevents immediate scale-in.
A threshold of less than 35%35\% CPU Percentage.
To prevent immediate scale-in (flapping), the scale-in threshold must be lower than the post-scale-out average CPU of 37.5%37.5\%. A threshold of 35%35\% prevents the scale-in rule from triggering at 37.5%37.5\%, whereas a threshold of 50%50\% would trigger it immediately.

Key Concept

Azure App Service Autoscale Tiers and Flapping Prevention
Rate this question

Topics

Scale Azure App Service Web Apps

More questions from this topic