Examkin

Question

Difficulty: HardPerformance and Scalability Optimization

A fleet of 10,00010,000 delivery drones upload telemetry data concurrently at the top of every hour. During these peak windows, incoming traffic spikes from a baseline of 2,0002,000 requests per second (RPS) to over 100,000100,000 RPS within a 11-minute window. The architecture consists of an Application Load Balancer (ALB) that routes requests to an Auto Scaling group (ASG) of Amazon EC2 instances. The instances process the data and store it in an Amazon Aurora MySQL DB cluster. An analytical application also queries the Aurora cluster to display real-time coordinates, but experiences severe query latencies and timeouts during the peak traffic windows. Drones also experience connection timeouts and HTTP 503 errors.

Which architectural combinations should a solutions architect implement to optimize performance and scalability during the peak windows? (Select TWO.)

  1. Submit a support ticket to AWS to pre-warm the ALB to handle the expected instantaneous traffic spike, and configure Scheduled Scaling for the EC2 ASG to scale out before the top of the hour.Answer
  2. Configure Aurora Auto Scaling to dynamically provision Aurora Replicas based on reader CPU utilization, and configure the analytical application to use the Aurora reader endpoint.Answer
  3. C
    Enable Multi-AZ deployment on the primary database instance to automatically route analytical read queries to the standby instance during high-traffic spikes.
  4. D
    Configure the EC2 ASG to use a target tracking policy based on average CPU utilization, and set the scaling cooldown period to 1515 seconds to ensure rapid scale-out.
  5. E
    Deploy Amazon ElastiCache for Memcached with Multi-AZ replication to cache the analytical query results and minimize replication lag from the primary database instance.

Answer

Submit a support ticket to AWS to pre-warm the ALB and configure Scheduled Scaling for the EC2 ASG; and configure Aurora Auto Scaling with Aurora Replicas using the reader endpoint.
Pre-warming the ALB is required because Application Load Balancers scale gradually, and a 50-fold instantaneous spike will result in dropped packets and HTTP 503 errors. Scheduled scaling is required for the EC2 Auto Scaling group because reactive metrics (such as CPU utilization) would take too long to detect, provision, and bootstrap instances to handle the immediate load. Offloading analytical queries to Aurora Replicas via the cluster's reader endpoint removes the read burden from the primary DB instance, while Aurora Auto Scaling dynamically manages replica count based on actual load.

Step-by-Step Solution

1
Analyze the ingestion layer performance bottleneck.
The traffic spike increases by 50 times in under a minute, which exceeds the rate at which an Application Load Balancer (ALB) can automatically scale. Standard reactive scaling policies for the EC2 ASG will also trigger too late because instances take time to launch and bootstrap.
Requesting ALB pre-warming from AWS Support guarantees capacity is available immediately, while Scheduled Scaling ensures EC2 capacity is ready before the traffic spike begins.
2
Evaluate the database layer read scaling options.
The analytical application is causing high latency on the primary database instance because it reads from the same instance that handles ingestion writes.
By creating Aurora Replicas and routing analytical queries through the reader endpoint, the read workload is offloaded from the primary instance. Aurora Auto Scaling can then dynamically add replicas if read demand increases further.

Key Concept

Handling sudden flash traffic spikes using pre-warming and scheduled scaling, combined with horizontal read scaling of database workloads using Aurora Replicas and reader endpoints.
Rate this question

Topics

Performance and Scalability Optimization

More questions from this topic