Cost control during AZ outage question – not sure answer is correct

[tomw]

Hi,

This question below asks about a cost effective architecture to keep an application available during an AZ outage during peak load, and suggests the use of spot instances. I note that peak load in a region is probably about the same time that other customer facing applications are experiencing peak load, business hours.

When an AZ goes down the other AZs in the region typically get very busy and there is a good chance that all spot instances are terminated. Reserved instances with capacity reservations are filled first, then on-demand, then spot.

If it’s critical the application stays up then reserved instances with capacity reservations are required in each region. At worst on-demand instances should be used. Spot should not be used in these circumstances. You might get away with it in a huge region like us-east-1, but smaller regions like ap-southeast-2 you might not.

Question

https://portal.tutorialsdojo.com/courses/aws-certified-solutions-architect-professional-practice-exams/lessons/practice-exams-review-mode-5/quizzes/aws-certified-solutions-architect-professional-practice-exam-review-mode-set-1/

A web application is composed of an Application Load Balancer and EC2 instances across three Availability Zones. During peak load, the web servers operate at 95% utilization. The system is set up to use Reserved Instances to handle steady state load and On-Demand Instances to handle the peak load. Your manager instructed you to review the current architecture and do the necessary changes to improve the system.

Which of the following provides the most cost-effective architecture to allow the application to recover quickly in the event that an Availability Zone is unavailable during peak load?

Review Mode Set 1 – AWS Certified Solutions Architect Professional

[Kenneth-Samonte-Tutorials-Dojo]

Hi Tomw,

I understand that Spot instances are not a guarantee. However, there are several instances and pricing choices for Spot instances. If there are only small Spot instances that are available, you can spawn more of them to accommodate the peak traffic. The application will still always be running because you still have reserved instances to handle the steady-state load.

We have addressed this question before on this thread.

Spot instance is not a guaranteed instance type

BTW, we also have a few updates on this question, to properly reflect the correct wording of the scenario. You will see it reflected as soon as the question bank is updated

Regards,

Kenneth Samonte (Tutorials Dojo)

[tomw]

Thanks Kenneth. An application will probably stay available using spot fleets during an AZ outage, but they might not as capacity will be tight. If an application *has* to stay up then reserved capacity is important.

I would probably tweak the question / answers a bit to de-emphasize the AZ outage part, as people who’ve worked in AWS in the real world for some time are probably more likely to get this wrong than someone with less experience.