Soru

Zorluk: Çok zorScale Azure App Service Web Apps

An organization runs a CPU-intensive background job processing application on an Azure App Service Web App that currently uses the Basic (B2) pricing tier. During business hours, the processing queue grows rapidly, causing delays. You must implement an autoscaling strategy that automatically scales the application out up to a maximum of 88 instances when CPU utilization spikes, and scales in when demand drops. The solution must minimize monthly hosting costs and prevent autoscale flapping.

Which scaling configuration should you recommend?

  1. Scale the App Service Plan to the Standard (S1) tier. Configure a scale-out rule to increase the instance count by 11 when the average CPU Percentage is greater than 80%80\% for 1010 minutes, and a scale-in rule to decrease the instance count by 11 when the average CPU Percentage is less than 35%35\% for 1010 minutes.Cevap
  2. B
    Scale the App Service Plan to the Standard (S1) tier. Configure a scale-out rule to increase the instance count by 11 when the average CPU Percentage is greater than 80%80\% for 1010 minutes, and a scale-in rule to decrease the instance count by 11 when the average CPU Percentage is less than 65%65\% for 1010 minutes.
  3. C
    Scale the App Service Plan to the Basic (B3) tier. Configure a scale-out rule to increase the instance count by 11 when the average CPU Percentage is greater than 80%80\% for 1010 minutes, and a scale-in rule to decrease the instance count by 11 when the average CPU Percentage is less than 35%35\% for 1010 minutes.
  4. D
    Scale the App Service Plan to the Premium V3 (P1v3) tier. Configure a scale-out rule to increase the instance count by 11 when the average CPU Percentage is greater than 80%80\% for 1010 minutes, and a scale-in rule to decrease the instance count by 11 when the average CPU Percentage is less than 65%65\% for 1010 minutes.

Cevap

Scale the App Service Plan to the Standard (S1) tier. Configure a scale-out rule to increase the instance count by 11 when the average CPU Percentage is greater than 80%80\% for 1010 minutes, and a scale-in rule to decrease the instance count by 11 when the average CPU Percentage is less than 35%35\% for 1010 minutes.
The configuration using the Standard (S1) tier and a 35%35\% scale-in threshold is correct because the Standard tier is the lowest-cost tier that supports automated scale-out (up to 1010 instances). The 35%35\% scale-in threshold ensures that when scaling down from 22 instances to 11 instance, the CPU load of the remaining instance will not exceed the 80%80\% scale-out threshold, thereby preventing autoscale flapping.

Adım Adım Çözüm

1
Determine the minimum pricing tier that supports autoscale up to 88 instances.
Standard (S1) tier is selected. Basic tier does not support autoscale rules, and Premium tiers are more expensive and unnecessary for a maximum of 88 instances.
The solution must minimize monthly hosting costs while supporting autoscale up to 88 instances.
2
Analyze the autoscale thresholds to prevent flapping when scaling in.
The scale-in threshold must be configured such that the load on the remaining instances after scale-in does not immediately trigger a scale-out. Mathematically, for a scale-out threshold of 80%80\%, scaling down from 22 instances to 11 instance requires the scale-in threshold to be less than 40%40\% (since 2×39%=78%2 \times 39\% = 78\%, which is less than 80%80\%, whereas 2×65%=130%2 \times 65\% = 130\%, which is greater than 80%80\%, triggering an immediate scale-out).
Flapping occurs when a scale-in operation immediately causes the remaining instances to exceed the scale-out threshold, triggering a scale-out again.
3
Select the configuration option that satisfies both the tier and threshold criteria.
Standard (S1) tier combined with a scale-out threshold of 80%80\% and a scale-in threshold of 35%35\% satisfies all requirements.
This is the only option that uses the cheapest valid pricing tier and prevents autoscale flapping.

Anahtar Kavram

App Service Plan pricing tier capabilities and autoscale rule configuration to prevent flapping.
Bu soruyu puanla