HackerRank reduces EC2 costs by 50% while modernizing build pipeline

DoiT recommended a multi-step approach including optimizing Docker container startup time, running unit tests in parallel, and moving the build server to Amazon ECS with AWS Fargate. The FinOps team implemented PerfectScale for Spot to optimize Auto Scaling groups in non-production environments, enabling confident adoption of Spot Instances with fallback to on-demand when needed.

HackerRank, founded in California in 2012, enables more than 3,000 companies worldwide to hire skilled developers. With HackerRank's proprietary platform, employers can screen and evaluate candidates' programming skills through coding challenges, assessments, and technical reviews. The platform enables tech recruiters and hiring managers to objectively evaluate talent at every stage of the recruiting process. HackerRank also offers a community platform where developers can share knowledge, showcase skills, solve coding challenges, participate in contests, and collaborate with other developers worldwide.

DoiT identified that HackerRank's build-release pipeline introduced delays and lengthened release cycles. The solution included optimizing Docker container startup time, providing code samples and documentation for proof of concept, running unit tests in parallel, and moving the build server to Amazon Elastic Container Service with AWS Fargate. This approach provided greater elasticity and lower maintenance while streamlining the entire build-release process.

DoiT's FinOps team implemented PerfectScale for Spot to optimize Auto Scaling groups organizing EKS worker nodes in non-production environments. Using machine learning, the technology continuously monitors cloud usage to identify compute not covered by existing commitments. PerfectScale for Spot made best-practice recommendations to ASG configurations, supporting more instance types with similar attributes and including a fallback to on-demand feature when Spot capacity is insufficient.

The Pipeline Modernization project enabled HackerRank to release small changes more frequently with greater agility, aligning with Agile practices of incremental delivery. This resulted in faster, more reliable service to customers. The cost optimization efforts delivered a 24% decrease in on-demand EC2 expenses and nearly 50% reduction in on-demand EC2 spend for non-production ASGs, saving thousands of dollars without additional effort or risk.