Leveraging automation and alerting for sustainable cost control within a multi-region AWS environment
Challenge
Our customer in the Ad Tech industry had built their marketing platform from the ground up on AWS. Foundational components were centralized in a primary AWS region, but as the product expanded to additional regions and single-tenant capabilities were added, their cloud costs also began to grow.
What was once a simple system with a unified tech stack eventually became complex with very little visibility into how resources were being utilized.
Infrastructure costs were growing exponentially. Controlling their AWS spend before it got out of control quickly became a top priority. With product deadlines looming and resource availability tight, they decided to engage Protagona to identify areas for immediate savings and implement controls to prevent overspending going forward.
Solution
Protagona quickly identified and conducted several cost cleanup exercises to immediately reduce unnecessary spending. Tagging strategies were then defined and implemented within automation templates to assign clear ownership of cloud resources. Cost controls were then put in place to automatically identify orphaned or underutilized resources and notify owners to take action.
Database and EKS clusters were then evaluated for cost efficiencies. In both cases, clusters were modified to take advantage of horizontal auto-scaling. In addition, non-prod EKS clusters were updated to leverage spot instance nodes for additional cost savings.
Tech Stack
AWS EC2
AWS S3
AWS EKS
AWS RDS
AWS Lambda
Terraform
Jenkins
Prometheus
Datadog
Grafana
Timescale DB
Outcome
Cost Savings – Protagona identified cost savings improvements that will end up saving the client between 20% and 30% on their monthly AWS bill. As their platform continues to grow, cost controls implemented will continue to drive additional savings moving forward.
Better Observability – Monitoring and alerting were implemented to increase visibility into current cloud spend and automatically take action to clean up resources when appropriate. Increased visibility into resource utilization has led to right sizing exercises and more accurate capacity planning.
Culture – A consistent tagging strategy was implemented across AWS environments to assign clear ownership of resources. The client’s technology teams are now enabled to take full ownership of their product, including infrastructure cost optimizations and scaling improvements.