Evolving a Continuous Delivery Workflow
summary
Schibsted evolves a Spinnaker-based CD pipeline over a year from UI point‑and‑click to pipelines-as-code, building a custom Training Wheels service to apply manifests, then migrating to Spinnaker’s manifest-based Kubernetes integration. The platform serves ~100 sites and ~1,000 developers, runs about 400 production deployments per week, injects cluster-specific data via a deploy daemon, enforces namespace authorization, exposes commit/build metadata in the UI, and reports that templated, opinionated configs cut maintenance and speed up cluster migrations.
AI-written summary · about
chapters (11)
- 0:00Intro
- 0:28Our challenge
- 1:46Our mission: Promote CD and build infrastructure to support it
- 3:52Our golden path for deploying applications
- 4:34What is Spinnaker
- 8:28First iteration of kos deployments
- 9:58Training wheels
- 15:41Pipelines as Code - declarative pipelines
- 23:49FIAAS + Manifest based deployment
- 29:53Next steps
- 32:04Takeaways
related talks
31mHow Spinnaker is Shaping Delivery Excellence at SAP
43mAll Aboard: Transforming Software Delivery Using Spinnaker
20mEmpowering Continuous Delivery with Spinnaker: A State of th...
36mOrganizational Challenges of Implementing Continuous Delivery w/ Spinnaker
48mContinuous Delivery with Spinnaker
36mSpinnaker - Land of a 1000 Builds
33mPanel Discussion: Spinnaker Success - CD as a Service using Spinnaker
28mScale CI/CD in Your Enterprise Through An Internal Spinnaker as-a-Service Model