The scope of Rancher Support SLA is limited to when the Rancher management server is run on a Kubernetes cluster in an infrastructure provider where Kubernetes is installed using K3s or RKE.
Running Rancher on EKS, AKS (or any Hosted K8s Provider) is not a configuration we support.
The reason we do not support this configuration is that providers such as EKS/AKS/GKE do not allow enough control to the etcd database for the user to successfully operate and scale their Rancher environment. Any issues stemming from this configuration fall outside of our SLA and do not lend themselves to valid support tickets.
That said, in the event, you have somehow ended up with a Rancher install on a hosted k8s provider scenario, we always like to understand your context and motivations better and continue to help in any way we can, but on a best-effort basis. Is your motivation to run Rancher on a hosted k8s provider coming from a place of we want it to be easier managing RKE under a Rancher master deployment? Let us know.
It is on our roadmap to move away from requiring RKE to run HA. We are looking to simplify such that you only need nodes, a database, and a load balancer.
- Should this topic interest you for a conversation with the general Rancher community members, please feel free to add your own feedback to GH-16627 on GitHub.
- We will be glad to arrange a conversation for you with our Product Management team, if you would like to understand our reasons in detail and what we are looking into, to make it simpler to run Rancher HA.