Examkin

Question

Difficulty: MediumImplementing Auto Scaling and Fault Tolerance

A digital ticketing platform hosts high-profile concert ticket releases on AWS. During these releases, traffic surges from 100100 requests per second to over 50,00050,000 requests per second within a 22-minute window. The current architecture uses an Application Load Balancer (ALB) to distribute traffic to an Auto Scaling group (ASG) of Amazon EC2 instances in private subnets. The instances use a custom AMI with an application that takes 88 minutes to bootstrap and pass health checks. Outbound licensing checks are routed through a single NAT Gateway in a single Availability Zone.

During recent releases, users experienced HTTP 503503 Service Unavailable errors during the first few minutes of the surge. Additionally, the ASG over-provisioned instances, launching far more than needed before the initial scale-out instances could finish bootstrapping. Finally, a brief outage in the Availability Zone containing the NAT Gateway prevented licensing checks for all instances.

Which combination of actions should a solutions architect take to resolve these issues? (Select TWO.)

  1. Deploy a NAT Gateway in each Availability Zone with private subnet routing configured to use the local gateway, and request Application Load Balancer pre-warming before scheduled releases.Answer
  2. Set the Auto Scaling group instance warmup period to 540540 seconds to ensure instances complete bootstrapping before they contribute to the target tracking metrics.Answer
  3. C
    Deploy a NAT Gateway in a single public subnet to route outbound traffic for all Availability Zones, and rely on the Application Load Balancer's automatic scaling to handle the traffic surge without pre-warming.
  4. D
    Configure the Auto Scaling group default cooldown and instance warmup periods to 6060 seconds to allow the scaling policy to rapidly launch additional instances in response to sudden CPU spikes.

Answer

Deploy a NAT Gateway in each Availability Zone along with requesting Application Load Balancer pre-warming, and set the Auto Scaling group instance warmup period to 540540 seconds.
Deploying a NAT Gateway in each Availability Zone ensures that outbound routing is redundant and resilient to single-zone failures. Requesting Application Load Balancer pre-warming prepares the load balancer for massive traffic spikes that exceed standard automatic scaling rates. Configuring the instance warmup period to 540540 seconds (which is longer than the 88-minute bootstrapping time) ensures that the target tracking policy does not include warming-up instances in its metric calculations, preventing over-provisioning and scaling thrashing.

Step-by-Step Solution

1
Deploy a NAT Gateway in each Availability Zone and update the private subnet route tables to use the NAT Gateway in the same zone.
Establishes a fault-tolerant egress path where an outage in a single Availability Zone does not impact outbound traffic in other zones.
Resolves the single point of failure of the NAT Gateway and ensures licensing checks succeed during AZ outages.
2
Request Application Load Balancer pre-warming from AWS Support prior to the scheduled release events.
Pre-provisions ALB resources to handle the massive, immediate surge of traffic from 100100 to 50,00050,000 requests per second.
Prevents HTTP 503503 errors during the first few minutes of the surge because standard ALB scaling is too slow for instantaneous spikes.
3
Configure the Auto Scaling group's instance warmup period to 540540 seconds (99 minutes).
Excludes the CPU and request metrics of newly launched instances from target tracking averages until they are fully operational.
Prevents the Auto Scaling group from over-provisioning and thrashing, as the warmup period exceeds the 88-minute (480480-second) bootstrap time.

Key Concept

Designing for fault tolerance with multi-AZ NAT Gateways, pre-warming load balancers for flash traffic, and configuring appropriate Auto Scaling warmup periods to match instance bootstrapping times.
Rate this question

Topics

Implementing Auto Scaling and Fault Tolerance