SAP-C02 Question 509
Select 3Your organization runs a data analytics platform on AWS. The architecture includes Amazon Redshift for data warehousing, Amazon S3 for storage, and AWS Glue for ETL processes. Recently, the CFO has asked you to analyze and optimize the AWS bill for this platform. After reviewing the usage and costs, you notice the following:
- Amazon Redshift is running 24/7, but usage is significantly lower during weekends.
- AWS Glue jobs are configured with default worker types and are currently over-provisioned for many jobs.
- Large volumes of data are being retrieved from Amazon S3 to Amazon Redshift for queries.
What are the most effective actions you can take to optimize costs for this platform?
- A
Pause Amazon Redshift clusters during weekends to reduce compute costs.
- B
Switch Amazon S3 storage class to S3 Glacier Instant Retrieval for all data to reduce storage costs.
- C
Optimize AWS Glue jobs by selecting smaller worker types and adjusting parallelism based on job requirements.
- D
Enable Amazon Redshift Spectrum to query data directly from Amazon S3, reducing the need for data ingestion.
- E
Migrate data from Amazon S3 to Amazon EBS for faster retrieval with lower costs.
Show answer and explanation
Correct answers: A, C, D
Explanation
To optimize costs, you should pause Amazon Redshift clusters during periods of low usage to save on compute costs. Additionally, optimizing AWS Glue jobs by selecting appropriate worker types and parallelism ensures that resources are not wasted. Finally, using Amazon Redshift Spectrum to query data directly from S3 eliminates the need for unnecessary data transfers and storage duplication, further reducing costs. These actions align with AWS best practices for cost optimization while maintaining performance.
- A. Correct.
Pausing Amazon Redshift clusters during weekends when usage is low can significantly reduce compute costs, as Redshift charges are based on running clusters. This is an effective cost-saving strategy.
- B. Incorrect.
Switching to S3 Glacier Instant Retrieval is not ideal for active analytics use cases because it is designed for infrequently accessed data. This would increase retrieval costs and reduce performance for queries.
- C. Correct.
Optimizing AWS Glue jobs by selecting smaller worker types and fine-tuning parallelism ensures that resources match the workload requirements, effectively reducing costs without impacting performance.
- D. Correct.
Using Amazon Redshift Spectrum to query data directly from S3 avoids the need to load large volumes of data into Redshift, reducing both data transfer and storage costs while maintaining query performance.
- E. Incorrect.
Migrating data from Amazon S3 to Amazon EBS is not cost-effective for analytics platforms, as S3 offers lower costs for storage and is designed for scalability in data analytics workloads.