Highlights
ClusterCacheTracker can now better deal with situations where some workload clusters with a high numbers
of Machines are not reachable. This was achieved by introducing per-cluster locking and directly requeuing
whenever another reconciler worker is already trying to create a client for a specific workload cluster.
Additionally, this release contains a few bugfixes to consistently reconcile finalizers of MachineSets and
MachineDeployments and ownerReferences of Machines and MachineSets owned by MachineDeployments.
Changes since v1.2.5
✨ New Features
- ClusterCacheTracker: non-blocking per-cluster locking (#7544)
- add kubekey k3s boostrap and control plane provider (#7553)
- add kubekey provider (#7541)
🐛 Bug Fixes
- Add finalizer reconcile for Topology MachineSets and MachineDeployments (#7556)
- Adoption for stand-alone objects should ignore objects generated by MachineDeployments (#7539)
📖 Additionally, there have been 1 contributions to our documentation and book. (#7558)
Thanks to all our contributors! 😊