Continuous deployment at Skyscanner
summary
Skyscanner moves from slow, agent-based CI to container-first pipelines using Drone CI for ephemeral builds, Artifactory/ECR for images, and ECS blue‑green deployments with a Slingshot API layered on AWS SWF. The platform instruments traffic via an HAProxy sidecar into OpenTSDB/Grafana for automated checks, survives a container registry outage lesson, migrates 8,500 repos over 10 weeks, and achieves roughly 21,000 deploys/month by iterating incrementally rather than buying Kubernetes up front.
AI-written summary · about
chapters (15)
- 0:00Introduction
- 1:24Continuous deployment goals
- 3:07Defining CD models
- 5:05Initial challenges
- 8:56Scaling problems
- 12:02Implementing Drone CI
- 14:15Blue-green deployment
- 16:25Monitoring and visibility
- 18:42Slingshot platform
- 21:41CD benefits
- 27:06Container registry failures
- 29:11Source control migration
- 31:17Automated testing
- 32:22Future and impact
- 34:34Key takeaways
related talks
31mInception the container version
37mHow we scale DroneCI on demand
33mEvolving a Continuous Delivery Workflow
53mThe Road to Continuous Deployment: a Case Study
48mContinuous Delivery with Spinnaker
49mThe road to continuous deployment: a case study
52mContinuous Deployment to the Cloud using Spinnaker
48mScaling a Migration to CD at Airbnb
