A team focused on re-engaging dormant users utilized Amazon Web Services (AWS), including Amazon Lambda and Amazon Simple Email Service, to conduct rapid experiments and infrastructure iterations, leveraging data from CloudWatch and New Relic for monitoring. Initially using API polling for CloudWatch data, they faced challenges such as latency, incomplete metrics, and high costs, prompting a switch to CloudWatch Metric Streams and Terraform, which improved latency, reduced costs, and provided comprehensive metric data without custom namespaces. By implementing an infrastructure-as-code approach with Terraform, they managed to streamline the setup and maintenance of CloudWatch Metric Streams, allowing for better integration with New Relic and facilitating enhanced monitoring capabilities. The transition included setting up necessary AWS permissions, creating CloudFormation stacks, and updating dashboards and alerts to utilize the new streaming metrics, ultimately achieving more efficient monitoring and cost-effective data management.