Case Study: Lyft achieves up to 75% CI testing cost savings with Amazon Web Services (AWS) Spot instances

Lyft - Customer Case Study

Lyft, the fast-growing San Francisco rideshare company, runs a microservices infrastructure on AWS and uses Jenkins to orchestrate a continuous integration (CI) pipeline. Routine pre-production tests require large pools of compute resources, driving up daily infrastructure costs as builds and tests scale.

To cut costs, Lyft switched many CI jobs to Amazon EC2 Spot instances—making a four-line change to a Salt module that launches Spot capacity—and started using older-generation, lower-cost instances that can run more containers. The move yielded dramatic savings (up to ~90% on some CI processes and about 75% in monthly testing costs), while increasing throughput so tests finish faster.

Lyft

Matthew Leventi

Engineer, Lyft