Examkin

Soru

Zorluk: Çok zorOptimizing Compute and Storage Performance

An enterprise runs a high-performance batch analytics system on AWS. The ingestion tier uses an Application Load Balancer (ALB) to distribute incoming files to a fleet of memory-optimized Amazon EC2 instances in an Auto Scaling group (ASG). These instances download raw datasets from Amazon S3, process them, and write intermediate results to attached 1 TB1\text{ TB} `gp3` Amazon EBS volumes. The processed metadata is then written to an Amazon RDS for PostgreSQL database, which replicates to a read replica to serve real-time dashboard queries.

During the weekly batch run, the Solutions Architect identifies the following issues:
* The EC2 instances experience significant disk I/O wait times. The `gp3` volumes are configured with the default 3,000 IOPS3,000\text{ IOPS} and 125 MB/s125\text{ MB/s} throughput, but CloudWatch metrics show `VolumeThroughput` regularly hitting its limit.
* The RDS PostgreSQL database experiences high replication lag on the read replica, which impacts dashboard query accuracy.
* The incoming request volume spikes instantly at the start of the batch run, causing HTTP 503503 errors at the ALB before the EC2 fleet can scale out to handle the load.

Which two actions should the Solutions Architect take to resolve these bottlenecks and optimize system performance? (Choose two.)

  1. Modify the EBS `gp3` volume configurations to increase the provisioned throughput to 1,000 MB/s1,000\text{ MB/s} and scale the provisioned IOPS.Cevap
  2. Contact AWS Support to pre-warm the Application Load Balancer prior to the scheduled batch initiation windows.Cevap
  3. C
    Configure the real-time dashboard queries to target the secondary standby database instance in the RDS Multi-AZ deployment to offload read operations.
  4. D
    Configure the Auto Scaling group scaling policy with a cooldown period of 10 seconds10\text{ seconds} to accelerate instance deployment when spikes occur.
  5. E
    Deploy an Amazon ElastiCache for Memcached cluster with multi-AZ replication enabled to cache database query results and reduce replica load.

Cevap

To optimize the system, the EBS gp3 volume throughput should be scaled to 1,000 MB/s, and AWS Support should be contacted to pre-warm the Application Load Balancer before the batch run.
Increasing the provisioned throughput of the gp3 volumes to 1,000 MB/s directly resolves the storage bottleneck, as the default throughput of 125 MB/s is the limiting factor for the batch workload. Contacting AWS Support to pre-warm the Application Load Balancer ensures that the ALB is pre-scaled to handle the sudden, massive spikes in incoming traffic at the start of the batch run, preventing HTTP 503 errors.

Adım Adım Çözüm

1
Analyze the EBS I/O bottleneck.
Identify that the default 125 MB/s throughput on the gp3 volumes is being saturated by the batch writes.
gp3 volumes default to 3,000 IOPS and 125 MB/s throughput, but they can be scaled up to 1,000 MB/s throughput independently of capacity.
2
Address the immediate request drops at the ALB.
Determine that the instant traffic spikes exceed the ALB's automatic scaling rate, resulting in HTTP 503 errors.
Pre-warming the ALB by contacting AWS Support ensures pre-allocated capacity is ready before the batch run begins.
3
Evaluate the database replication and scaling configuration.
Confirm that Multi-AZ standby instances do not accept read traffic, and decreasing Auto Scaling cooldowns below the bootstrap duration creates resource churn.
Ensures that incorrect scaling and standby utilization patterns are avoided.

Anahtar Kavram

Optimizing compute scaling limits, EBS gp3 performance configurations, and handling flash traffic on Application Load Balancers.
Bu soruyu puanla

Konular

Optimizing Compute and Storage Performance