Unconference Session: Reliable K8s Operator: How to Survive Eventual Consistency

Don't miss out! Join us at our next KubeCon + CloudNativeCon events in Mumbai, India (18-19 June, 2026), Yokohama, Japan (29-30 July, 2026), and Shanghai, China (8-9 September, 2026). Connect with our current graduated, incubating, and sandbox projects as the community gathers to further the education and advancement of cloud native computing. Learn more at https://kubecon.io

Unconference Session: Reliable K8s Operator: How to Survive Eventual Consistency

Unconference Lead: Marek Siarkowicz

We've seen it before: the daemonset controller running wild, creating duplicate pods on a single node. Or the scheduler stops scheduling for hours, while logs and metrics show nothing amiss. This isn't a bug in the controller logic; it's a symptom of a deeper problem.

The controller or operator makes logical decisions based on an outdated world view. In K8s watch-based, eventually consistent world, long delays break controllers. The question here is: how far behind is the controller's cache from the apiserver's state?Fortunately, recent proposals in K8s tackle this problem.

During the session we want to start discussion on how we integrate those capabilities into operators allowing them to that degrade gracefully, and establish new design patterns that will make the CNCF ecosystem more observable and reliable for everyone.

Видео Unconference Session: Reliable K8s Operator: How to Survive Eventual Consistency канала CNCF [Cloud Native Computing Foundation]