The Oregon Trail to Kubernetes [I]
summary
Small team migrates from Google Compute Engine to GKE running 2000 CPUs, abandons YAML templating and implements Kubernetes resources as client-go-based CLI with compiled type checking, a deployment-set abstraction for 1-to-N replicas, and a staging updater integrated with Slack. Uses pod preStop hooks for long-running workflows, daemonset log shippers into ELK, flattens pod metric names for carbon, overcommits CPU to exploit the scheduler, and still defers Elasticsearch off-cluster.
AI-written summary · about
chapters (10)
- 0:00Start
- 3:01The Lytics Trailhead
- 4:50Loading the Wagon Goals
- 9:19Compute Resource Massacre
- 10:31On the Trail
- 17:35Rigging a CLI
- 20:33Blizzards of Kubernetes
- 22:56Handyman's Corner
- 26:58Interactive Operators
- 32:46The Homestead
related talks
35mKubernetes journey
31mMigrating Legacy Infrastructure to Kubernetes: Olark’s Lessons From the Trenches
02h 18mDevelop and Deploy to Kubernetes like a Googler
32mMigrating 2 million CPU cores to Kubernetes
37mBalancing Power and Pain: Moving a Startup From a PaaS to Kubernetes
44mMigrating a Monolith to Kubernetes
35mNavigating the Kubernetes Odyssey: Lessons from Early Adoption and Sustained...
24mFrom Lift-and-Shift to K8s: A Two-Phase DevOps Modernization