Session
The Cache That Outlives the Pod: Fast Container Builds on Ephemeral Kubernetes CI Runners
Ephemeral Kubernetes CI runners give every job a fresh pod, and are widely used with Actions Runner Controller (ARC), Tekton, and GitLab Runner. The catch? Every job starts from scratch, pulling base images, installing dependencies, and rebuilding Docker layers, even when little has changed.
In this session, we’ll explore how to speed up builds without giving up ephemeral runners. Using GitHub Actions and ARC as an example, we’ll demonstrate how a shared, long-running BuildKit service backed by persistent storage lets runners reuse build cache, cutting build times from minutes to seconds. We’ll compare cold and cached builds, then take down the builder mid-run to see how the cache holds up.
People will learn why persistent runners aren’t the only answer, how to adopt this approach with minimal workflow changes, and what to consider in production. We’ll cover the security & operational trade-offs, like mTLS, cache poisoning, and challenges of scaling beyond a single replica.
Please note that Sessionize is not responsible for the accuracy or validity of the data provided by speakers. If you suspect this profile to be fake or spam, please let us know.
Jump to top