Scaling and upgrades

Worker nodes

Additional workers of your cluster size are ordered in the customer portal and added by us; pods are scheduled onto them automatically. Removing nodes works the same way, with a drain first.

Upgrades

We follow upstream RKE2 releases and upgrade quarterly in announced maintenance windows. Worker upgrades drain and reschedule pods, so keep replicas above one for anything that must stay up.

Backups

We back up etcd (the cluster state). Application data on persistent volumes is yours to back up; ZERO-Z3 with Object Lock is a fitting target, see Veeam or any Kubernetes backup tool that writes to S3.