Emir Beganović and Luka Stout, Booking.com
Your CI hosts show plenty of idle CPU, yet a job you gave 4 cores behaves like it has 2. Where did they go?
At Booking.com we hit exactly this running ~10 million CI jobs a month on a container-based fleet. The bottleneck wasn't hardware but rather a combination of CFS throttling on high thread-count servers. This talk walks through what we did instead: replacing shared-kernel containers with fully isolated ephemeral KVM microVMs on the same hardware. This bought us ~3× more effective CI capacity, better p95 latency, and a hard isolation boundary, while preserving exact developer experience with zero CI syntax changes for users.
Using four laws as a guide: Little's Law, Amdahl's, Universal Scalability, and Hyrum's, you'll leave with a concrete way to tell when a shared scheduler has stopped fitting your workload, and why partitioning can beat sharing for CI workloads.
Emir Beganović is a Staff SRE at Booking.com in Developer Experience, working on the build and CI platform that thousands of engineers depend on every day. His career spans startups, scaleups, and enterprises across four countries. He's been a long-time Linux user and community contributor and established an Internet Exchange Point, and these days he's happiest at the boundary between CI systems, platform engineering and developer experience.
Luka Stout is a Senior SRE at Booking.com, where he works on making thousands of engineers more productive through better CI/CD, testing, and observability. He came to SRE from the data and ML world, with 6+ years across companies like MessageBird, where he cut CI/CD time by 75%. He holds an MSc in Artificial Intelligence from the University of Amsterdam.
