16:30: Doors open, find a seat or network :)
17:00: Welcome by Cloud Native Aalborg
17:05: Welcome by Centrica
17:15: "Flexible node provisioning with Karpenter" by Valer Orlovsky, Cloud Platform Engineer @Centrica
18:00: Break & food
18:30: "GitOps: Catch 22 - why state is not always what you think" by Henrik Høegh, Platform Engineer @Lunar
20:00: Thank you and good night!
"Flexible node provisioning with Karpenter" by Valer Orlovsky, Cloud Platform Engineer @Centrica
Karpenter is a new autoscaler that aims to make provisioning of Kubernetes worker nodes more efficient and cost-effective, by doing away with node groups. Instead, it manages the lifecycle of individual machines directly based on flexible configuration in form of CRDs. The project was created by AWS and is open-source. As of now, only the AWS provider has been implemented, however, the ambition of the core project is to be cloud-agnostic, which is also reflected in Karpenter's open design.
We have recently migrated some of our clusters from Cluster Autoscaler to Karpenter and are so far quite pleased with the results. In this talk, you will learn about how Karpenter works, how it is different from Cluster Autoscaler, and how it's making job of our Cloud Platform Team at Centrica Energy Trading easier.
Valer Orlovsky is a Cloud Platform Engineer, building the cloud infrastructure and developer platform at Centrica Energy Trading.
"GitOps: Catch 22 - why state is not always what you think" by Henrik Høegh, Platform Engineer @Lunar
GitOps can conceptually be hard to grasp, but are you using it correctly? Maybe you are only beginning to think about implementing it?
Defining the deference between declarative and imperative is important, when trying hold the state of truth.
Should pods defined in git or not?
Join this talk, to find out more about why GitOps done right, is essential for anyone implementing GitOps in their organisation.
Henrik Høegh is working as a Platform Engineer @Lunar by day and a speaker by night, he's been in the CNCF game for a long time and knows a thing or two about Kubernetes.