Question

Difficulty: MediumConfigure Azure App Services

An administrator manages an Azure App Service web app named WebApp1 that currently runs on a Basic (B1) App Service plan. The administrator needs to configure WebApp1 to scale out automatically based on CPU utilization. The configuration must prevent instance thrashing, where the system frequently and repeatedly adds and removes instances. Which of the following actions should the administrator perform? (Select TWO.)

  1. Scale up the App Service plan to the Standard (S1) pricing tier.Answer
  2. Configure autoscale rules with a scale-out threshold of 80% CPU and a scale-in threshold of 30% CPU.Answer
  3. C
    Scale up the App Service plan to the Shared (D1) pricing tier.
  4. D
    Configure autoscale rules with a scale-out threshold of 80% CPU and a scale-in threshold of 75% CPU.

Answer

Scale up the App Service plan to the Standard (S1) pricing tier, and configure autoscale rules with a scale-out threshold of 80% CPU and a scale-in threshold of 30% CPU.
To support autoscale, the App Service plan must be scaled up to at least the Standard (S1) tier since the Basic (B1) tier only supports manual scaling. Additionally, to prevent instance thrashing, the scale-in threshold must be set significantly lower than the scale-out threshold (e.g., 30% CPU for scale-in and 80% CPU for scale-out) so that the addition of an instance does not immediately trigger a scale-in.

Step-by-Step Solution

1
Determine the required App Service plan tier.
Identify that the current Basic (B1) plan only supports manual scaling up to 3 instances and does not support autoscale rules. To enable autoscale, the plan must be scaled up to at least the Standard (S1) tier.
Azure App Service autoscale rules require the Standard, Premium, or Isolated pricing tier.
2
Determine the correct autoscale threshold configuration to prevent thrashing.
Select thresholds that have a wide margin between scale-out (80% CPU) and scale-in (30% CPU) metrics.
If the scale-in and scale-out thresholds are too close (such as 80% and 75%), the decrease in average CPU load after scaling out will immediately trigger a scale-in, causing continuous, inefficient scaling cycles.

Key Concept

Configuring App Service autoscale rules and pricing tier capabilities.
Rate this question