Kubernetes from Dev to Prod at GoEuro
summary
Company migrates from legacy infra to Kubernetes across dev, CI, QA and prod in about four months, scaling from ~50 to 250 engineers and from 10 to 50 services with releases jumping from 10 coordinated to 150 independent releases/week. Development uses cloud single-node hyper VM (Minikube-like), repos include ops folders, Helm charts are rendered via an internal y8s service over HTTP, CI uses GitLab-style YAML with Jenkins agents on hyper VMs and preview VMs per PR, and a lightweight Theta proxy enforces namespace limits, annotations, whitelisted APIs and compliance before deploy.
AI-written summary · about
chapters (11)
- 0:00Start
- 0:37Transition Journey
- 3:58Development
- 11:59Composition
- 12:59Patterns
- 17:40Value additions: Cluster services
- 18:33Value additions: Sidecar and Regional services
- 20:05Value additions Annotation driven
- 20:54Landscape
- 21:27What we learned?
- 30:40Stepping Forward
related talks
35mKubernetes the easy way
45mLearning How to Pronounce Kubernetes to Production in 3 Months!
40mFrom dev to prod. One Year of Using Kubernetes in Production
35mKubernetes journey
29mA history of GOV.UK's infrastructure: the long voyage to Kubernetes
22m10 Kubernetes is hard!
44mMigrating a Monolith to Kubernetes
41mKubernetes — A Year in Production