This release packages the kura runner-cache storage stack directly into the CAPI Scaleway Helm chart, so the required scw-local-nvme StorageClass is now installed automatically whenever the kura fleet is enabled. You no longer need to apply a separate manual bootstrap manifest to provision the local-path storage that backs runner-cache PVCs.
- Automatic cache StorageClass: Enabling
kuraFleet.enablednow deploys thescw-local-nvmeStorageClass and its local-path provisioner alongside the rest of the fleet, preventing clusters from coming up without the storage the cache requires. - Reliable cold deployments: The local-path provisioner controller is no longer pinned to the elastic-metal cache node pool, so Helm upgrades do not hang waiting for a cache node during initial deployments.