Flo Health, a female health app with over 70 million monthly active users, has explained best practices for operating a large-scale app, optimized using Amazon DynamoDB, and achieving 60% cost efficiency.
To build a reliable and highly scalable database, DynamoDB was chosen to support the dynamic, read-heavy workload required for the 'Stories' feature, designed to curate interactive health education programs for women's health and improve user engagement.
Flo Health utilized AWS Well-Architected Framework (WAL) by performing an assessment of DynamoDB implementation, having an understanding of capacity mode analysis, and utilization of reserved capacity.
The company reduced costs by 20% by implementing Time-to-Live (TTL) and using reserved capacity, which reduced Write Capacity Units (WCU) usage by threefold.
Flo's data optimizations included datatype size optimization, whereby using numeric datatype instead of string datatype reduced the size of attribute across millions of items leading to saving substantial storage.
Flo also switched from string story ID to numeric ID leading to nine times more compact data storage, and implemented TTL for archiving or automatically expiring data that was no longer needed.
To enhance system efficiency and reduce costs, Flo implemented two key access optimizations: dirty checking and group update requests.
Overall, Flo's optimization of DynamoDB implementation improved system performance and cost efficiency, reducing provisioned WCU by 60%, lowering write activity and sustaining scalability.
By utilizing the AWS Well-Architected-Lens, similar results can be replicated to optimize cloud environments for scalability and cost-efficiency.
Flo Health's technical team is composed of Aleksej Klebanskij, Maksim Koutun, Maksim Ponasov, Lee Hannigan, and Mladen Trampic.