CNCF-hosted Co-located Events North America 2024

Everyone wants a judicial use of resources - and serverless computing makes it happen by limiting the idle resources. While services with clear functional use cases are developed as one, more often than not such benefits are not clear - such as unclear or changing application usage patterns. Similarly several historical services could now be suited for serverless functions based on they way they are actually used. However they are rarely investigated because their impact remains opaque - a lack of cost/performance benefits and whether re-architecture effort is justified. In this talk, the Yash will showcase a case study about how they went researching for serverless candidates in their real world systems and how surprising the results were. He will also showcase the impact and key learnings - how turning even lightweight microservices into serverless functions helped saved significantly on infra costs AND, also improving the peak load performance of several unrelated services!