Case Study: Pinterest achieves faster, more efficient builds with Kubernetes

Pinning Its Past, Present, and Future on Cloud Native

Pinterest, a cloud-native company running on AWS with more than 1,000 microservices and hundreds of thousands of data jobs, needed a faster, more consistent developer experience as its infrastructure became increasingly complex. To address this, Pinterest evaluated orchestration platforms and chose Kubernetes to help standardize packaging and improve the path from idea to production, starting with Jenkins workloads.

With Kubernetes, Pinterest migrated Jenkins Master to run natively on Kubernetes and used a shared cluster for builds, gaining elasticity, demand-based scaling, and improved failover policies. The vendor, Kubernetes, helped Pinterest reduce build times, reclaim over 80% of capacity during non-peak hours, and cut Jenkins cluster instance-hours by 30% per day compared with the previous static setup.

Pinterest

Micheal Benedict

Product Manager for the Cloud and the Data Infrastructure Group