01
Business needs
Claritas, a US data-driven marketing firm, faced a critical challenge with a soaring AWS cloud bill, reaching $244K monthly. The escalating costs, primarily from Amazon S3, Amazon Redshift, Amazon RDS, and Amazon OpenSearch, raised budget concerns. To address this, Claritas aimed to optimize costs through strategic measures and automation and align their cloud expenditure with budgetary targets.
The marketing firm wanted to:
- Implement best practices to optimize selected services to regain control over escalating costs.
- Implement and configure lifecycle and archival processes for object storage within the Amazon S3 feature.
- Plan and build automation mechanisms to identify and shut down unused resources, optimizing capacity reservation.
- Streamline Amazon Redshift and Amazon RDS usage to ensure efficient resource utilization.
- Identifying and eliminating unused resources through strategic planning and automation.
Achieved 35.8% monthly savings in Amazon S3 and 40.9% in Amazon Redshift costs
02
Solution
To counteract the challenges posed by escalating AWS costs, Impetus adopted a multi-faceted approach, employing a range of strategies to enhance cost efficiency and resource management:
- Used Amazon S3 storage lens and storage class analyzer for analyzing data in different buckets and applied life cycle policies to archive or migrate data to lower storage classes
- Committed using reserved instances for Amazon RDS, Amazon Redshift, and Amazon OpenSearch after estimating the resource requirements.
- Implemented automated pause/resume for Amazon Redshift clusters in non-prod accounts during off-business hours and weekends.
- Used Amazon CloudWatch to identify unused resources and deleted them after stakeholder’s confirmation.
- Increased CSP coverage to 80% based on utilization report.
- Downscaled Amazon Redshift cluster from 16 to 12 nodes based on Amazon CloudWatch metrics after workloads were migrated to Databricks.
Highlights
- Deleted 400 TB Databricks temporary data from Amazon S3 using lifecycle policies.
- 280 TB of backup data deleted, and 186 TB of data archived to Glacier
- Migrated 1.2 PB of data to intelligent tiering out of which 1 PB of data moved to Infrequent access.
- Updated Amazon RDS volumes to gp3 for improved performance.
- Migrated Amazon EBS volumes for Databricks workers from gp2 to gp3.
- Committed to use 80% reserved instances for Amazon RDS, Amazon Redshift, and Amazon OpenSearch.
AWS stacks used: Amazon S3, Amazon Redshift, Amazon OpenSearch, Amazon CloudWatch, Amazon EC2, Amazon RDS
03
Impact
The implemented strategies and optimizations had a profound impact on Claritas’ cloud expenditure, resulting in substantial cost reductions across key AWS services:
22.5% monthly reduction in the overall cloud bill, significantly enhancing cost efficiency.
- 35.8% monthly reduction in Amazon S3 costs, optimizing storage expenses.
- 40.9% monthly reduction in Amazon Redshift costs, enhancing efficiency in data warehousing.
- 36% monthly reduction in Amazon RDS costs, improving database cost-effectiveness.
- 25% monthly reduction in Amazon OpenSearch costs, enhancing search and analytics efficiency.
- 7% monthly reduction in Amazon EC2 costs, optimizing compute resources.
These impactful reductions demonstrate the successful alignment of Claritas’ cloud expenditure with budgetary targets and underscore the efficacy of the implemented cost optimization strategies in achieving sustained financial benefits.