Using Bazel for Fast, Correct Docker Deployments
summary
Bazel uses rules_docker plus Kubernetes-aware rules to turn build incrementality into tiny deploys: layered images split a 40 MB gRPC runtime from a 400 KB app so pushes shrink by two orders of magnitude, pushes and redeploys hit four seconds for a single change across five deployments, and image digests get inlined into JSONnet/Kubernetes templates. A Jenkins job is generated from JSON → XML, supports canary/dev groups, Slack notifications and audit history, and a one-line BUILD target wires CI to production-safe rollouts via kubernetes_object and ks_objects.
AI-written summary · about
chapters (6)
- 0:00Introductions and overview
- 1:25Evolution of building images
- 5:36Deploying with Bazel
- 14:18Live demonstration
- 22:47Automation at Databricks
- 29:02Scaling and Jenkins integration
related talks
40mStreamlining Kubernetes Application CI/CD with Bazel
07mLightning Talk – Bazel & Kubernetes: Develop like it's production
10mBuilding 1300 Container Images in 4 Minutes
15mReproducible research using Bazel
08mLightning Talk – Speedy Scala Builds with Bazel at Databricks
31mGradle to Bazel + Q&A
26mBuilding Go with Bazel
01h 01mIntroduction to Bazel