Upgrading Kubernetes sounds easy — until you’re operating a large fleet of production clusters, each with different workloads, requirements and complexity. This session will empower you to automate Kubernetes upgrades across a multi-tenant platform — with zero downtime, built-in observability, and automated verification.
This session will also cover lessons learned from building a custom Kubernetes controller using controller-runtime, building orchestration pipelines for Kubernetes upgrades, and maintaining service availability, even for singleton workloads that do not allow any disruption.

The audience will take away:

* Design patterns for safe, repeatable Kubernetes upgrades
* Integration of open-source tools like controller-runtime, and Spinnaker
* How to shift upgrades from a scary event to a routine, automated process