Scavenging for Reusable Code in the Kubernetes Codebase
summary
Replaces a custom pod-by-IP scanner with Kubernetes primitives, showing how a throwaway list-and-scan becomes an indexer backed by client-go cache.Indexer and a ListWatch/reflector then simplified further with shared informers, giving real caching, reconciliation and a sync check instead of arbitrary sleeps. Uses grep/google/godoc and throwaway go scripts and verifies behavior with the fake client test fixture. Also adopts work queues for TLS certificate distribution rate limiting; the replacements reduce code and work in production.
AI-written summary · about
chapters (5)
related talks
20mUsing Kubernetes API Effectively with Golang
35mProgramming Kubernetes with the Go SDK [I]
31mIt's all about reconciliation - anatomy of a kubernetes controller
35mThe Oregon Trail to Kubernetes [I]
27mPragmatic Pod Patterns: Leveraging sidecar containers in Kubernetes
46mSimplifying Scalable Cloud Application Development with Go
27mKubernetes The Fun Way
41mBuilding Apps in Kubernetes: The Unforgiving Sea of...