How Kubernetes CRD Versioning Breaks Controller Upgrades
from DevOps Daily with Fexingo: CI/CD, Kubernetes, and Modern Software Operations Hosted by Fexingo
Episode 86 of DevOps Daily with Fexingo dives into a subtle but painful Kubernetes trap: Custom Resource Definition (CRD) versioning. When you update a CRD's schema to add a new field or change a validation rule, existing Custom Resources stored in etcd may not automatically convert to the new version—and controllers that watch the old version can crash or miss updates. We walk through a real incident at a mid-sized SaaS company where a rollback attempt actually made things worse, because the old controller couldn't parse the new CRD storage version. You'll learn why the `storage` flag matters, how conversion webhooks can save you, and one simple check to add to your CI/CD pipeline before deploying CRD changes. #Kubernetes #CustomResourceDefinitions #CRDVersioning #ControllerUpgrades #etcd #KubernetesOperators #ConversionWebhooks #DevOps #SiteReliabilityEngineering #KubernetesRollback #CI/CD #CloudNative #KubernetesStorage #ControllerManager #APIExtensions #KubernetesTroubleshooting #FexingoBusiness #BusinessPodcast Keep every episode free: buymeacoffee.com/fexingo
Embed this episode
No transcript yet.