Baixar clusterctl 1.13.0 per a Windows

👌 Kubernetes version support
- Management Cluster: v1.32.x -> v1.35.x
- Workload Cluster: v1.30.x -> v1.35.x
More information about version support can be found here
Highlights
CAPI v1.13 is a release focused on stability, reliability and performances:
- Bumped to Go 1.25, controller-runtime v0.23, k8s.io/* v0.35, controller-gen v0.20
- KCP now allows to recover from different kind of failures as well as from multiple failures:
- Increased tolerance to unexpected state of control plane components (accidental node deletion, manual removal of etcd members, accidental deletion of kubeadm control plane labels, and more)
- Better visibility on common symptoms of kubeadm join errors (failed to promote etcd member, failed to apply kubeadm control plane labels and taints)
- Multiple improvements to improve controllers memory and CPU footprint, performance and stability at scale
Note: Cluster API is only supported on conformant Kubernetes Clusters and contract-relevant provider resources (e.g. InfraCluster) have to be implemented as CRDs (i.e. not via an aggregated apiserver).
Other changes
- The implementation of the Taint propagation proposal is now completed
- Runtime Hooks are now embedding v1beta2 Cluster types
- Machine now reports info about machine placement in the status.failureDomain field
- rolloutAfter for control plane and MachineDeployments can now be controlled from Cluster.spec.topology
- MachineWaitForVolumeDetachConsiderVolumeAttachments feature graduated to GA
- PriorityQueue feature graduated to beta and it is now enabled by default
- ReconcilerRateLimiting feature graduated to beta and it is now enabled by default
- Starting from this release ReconcilerRateLimiting feature also requires PriorityQueue to be enabled. This ensures that ReconcilerRateLimiting works consistently with controller runtime ExponentialBackoff.
Deprecation and Removals Warning
- v1alpha3 and v1alpha4 apiVersions have been removed
- Reminder: v1beta1 is on track to be unserved in CAPI v1.16
- Reminder: Provider should start implementing the v1beta2 contract ASAP.
For additional details for providers, please take a look at Cluster API v1.12 compared to v1.13.[GitHub issue #REPLACE ME](REPLACE ME)
Changes since v1.12.0
📈 Overview
- 279 new commits merged
- 10 breaking changes ⚠️
- 30 feature additions ✨
- 39 bugs fixed 🐛
📝 Proposals
- Core: Update autoscaling from zero enhancement proposal with node labels and taints configuration clarification (#13308)
⚠️ Breaking Changes
- API/Runtime SDK: Bump Runtime Hooks v1alpha1 types to core CAPI v1beta2 (#13200)
- API: Remove v1alpha3 + v1alpha4 apiVersions (#13199)
- CABPK: Remove deprecated --cluster-concurrency flag in CABPK (#13288)
- CAPD: Mark docker resources as deprecated (#13565)
- ClusterCache: Remove deprecated ClusterCache GetClientCertificatePrivateKey (#13156)
- ClusterClass: Remove deprecated ShouldSkipImmutabilityChecks (#13291)
- clusterctl: Remove deprecated --disable-grouping clusterctl describe cluster flag (#13289)
- clusterctl: Remove deprecated provider CRD migration from clusterctl upgrade (#13290)
- Dependency: Bump to controller-runtime main & controller-tools v0.20 (#13159)
- Misc: Remove deprecated ParseMajorMinorPatchTolerant / ParseMajorMinorPatch (#13292)
✨ New Features
- API: Backport newly introduced v1beta2 API fields to v1beta1 (#13455)
- CABPK: Optimize cache configuration of CABPK & standardize cache/client setup (#13407)
- CAPD: Implement DevMachinePools (#13346)
- CAPD: Improve CAPD wait for multi-user target (#13514)
- CAPD: Support externally managed LB (#13362)
- Cluster: Only set ownerReference on InfraCluster/ControlPlane when Topology is defined (#13332)
- ClusterCache: Add ClusterFilter to ClusterCache Options (#12665)
- ClusterClass: Implement support for taints (#13192)
- clusterctl: Allow overriding image name in clusterctl config (#13014)
- Core: Optimize cache configuration of core CAPI (#13488)
- Dependency: Bump Go to v1.25.5 (#13164)
- Dependency: Bump to controller-runtime v0.23 (#13245)
- Devtools: Add a flag to skip image preloading (#13143)
- KCP/CABPK: Allow diskSetup to include partition layout (#11634)
- KCP/MachineDeployment/Cluster: Add rolloutAfter to cluster.spec.topology (#13391)
- KCP: Allow remediation of multiple failures in KCP (#13352)
- KCP: Bump coredns/corefile-migration to v1.0.30 (#13282)
- KCP: Implement support for machine taints (#13181)
- KCP: Optimize cache configuration of KCP (#13460)
- Machine: Add Machine status.failureDomain (#13266)
- Machine: Promote MachineWaitForVolumeDetachConsiderVolumeAttachments feature to GA (#13293)
- Metrics/e2e: Add infra CRDs to dashboards, improve/fix dashboards, enable native histograms (#13354)
- Misc: Disable DWARF and symbol table to decrease binary/image size (#12856)
- Misc: Enable PriorityQueue per default (#13171)
- Misc: Promote ReconcileRateLimiting to beta (enabled per default) (#13373)
- Misc: Tune controller concurrency and cache timeout for scale (#13496)
- Release: Prepare main branch for v1.13 (#13174)
- Testing: Add example / debug suite for envtest (#13453)
- util: Extend conversion tests to cover the "no spec" case (#13409)
- util: Refresh cache entries on cache hit in SSA caches (#13459)
🐛 Bug Fixes
- API: Fix v1beta1 ControlPlane contract to handle .status.initialized correctly (#13186)
- CAPD: Fix "Failed to exec DockerMachine bootstrap" errors in CAPD (#13447)
- CAPD: Remove finalizers during deletion if ownerRef was never set for cluster controllers (#13239)
- CAPIM: Extend CAPD in-memory backend to set CP taint to fix scale tests (#13187)
- CAPIM: Fix in-memory watch unit test (#13464)
- CI: Bump trivy to v0.69.2 to fix CI (#13387)
- Cluster: Fix panic in Cluster conversion (#13383)
- ClusterClass/KCP/MachineSet/MachineDeployment: Mitigate managedFields apiserver issue for SSA (#13338)
- ClusterClass: Allow adding spec via ClusterClass JSON patches (#13225)
- ClusterClass: Do not overwrite global http.DefaultClient TLSConfig (#13058)
- ClusterClass: Fix upgradePlan computation in GetUpgradePlanFromClusterClassVersions (#13463)
- Dependency: Bump dependencies to fix CVE GO-2026-4394 (go.opentelemetry.io/otel/sdk pkg) (#13372)
- Dependency: Fix CVE-2026-39883 (#13573)
- e2e: Add wait-resource-versions-{become,remain}-stable intervals to e2e config (#13263)
- e2e: Clean up namespace in e2e tests (#13233)
- e2e: Fix cross-ns scale test (#13309)
- e2e: Fix goroutine & memory leak in inmemory provider (#13361)
- e2e: Fix node-label calculation in test extension & fix upgrade test (#13356)
- e2e: Fix WaitForMachinesReady interval (#13050)
- e2e: Increase reconcile timeout for KCP & DockerMachine (#13093)
- e2e: Only retry creating objects that failed (#13265)
- e2e: Tolerate NotFound errors during Namespace deletion in scale test cleanup (#13439)
- KCP/MachineSet: Preserve existing object names for backward compatibility with pre-v1.7 in-place updates (#13124)
- KCP: Grant delete permissions to Secrets. (#13070)
- KCP: KCP deletion should tolerate missing InfraTemplates (#13562)
- KCP: Use errors.Errorf instead of errors.Wrapf since the err is nil (#13486)
- Machine/MachineSet/MachineDeployment: Fix UpToDate calculation for rolloutAfter (#13404)
- MachineSet: Use MachineSet template values in completeMoveMachine for in-place updates (#13059)
- Misc: Fix bug while setting status for deprecated fields (#13336)
- Runtime SDK: Improve client cert/key rotation of the RuntimeSDK client (#13213)
- Testing/e2e: Fix unit test flakes, improve clusterctl download error in e2e tests (#13045)
- Testing: Fix flaky by waiting for CRD finalizer processing (#13470)
- Testing: Fix flaky TestClusterReconciler unit test (#13180)
- Testing: Fix TestReconcile flake (#13255)
- Testing: Fix webhook envtest tests for Kubernetes < v1.35 (#13170)
- Upgrades: Remove conflicting rules field from aggregated ClusterRoles (#13490)
- util: Fix a panic in conditions.Delete method if the sources condition list is empty (#13048)
- util: Fix exponential backoff with ReconcilerRateLimiting (#13416)
- util: Fix patchHelper unit test flakes (#13412)
🌱 Others
- API: Deprecate custom Condition types (#13237)
- API: Introduce conversion.MarshalDataUnsafeNoCopy to avoid unnecessary memory allocations during conversion (#13402)
- API: Postpone date when we stop serving v1beta1 (#13394)
- API: Relax validation for Machine .status.addresses to maximum of 128 instead of 32 items (#13060)
- API: Relax validation for Machine .status.addresses to maximum of 256 instead of 128 items (#13395)
- CAPD/CAPIM: Implement .status.failureDomain for DockerMachine & DevMachine (#13286)
- CAPD: Implement pause for DockerMachinePool (#13445)
- CAPD: Improve condition if CAPD Machine is not yet ready for bootstrap exec (#13461)
- CAPD: Move RBAC for devmachinetemplates from main.go to controller.go (#13271)
- CAPD: Reduce verbosity of CAPD exec log (#13493)
- CAPD: Use select with time.After instead of time.Sleep (#13480)
- CAPIM: Fix inMemory watch (#13229)
- CI: Bump golangci-lint v2.7.0 (#13108)
- CI: Bump Kubernetes version used for testing to v1.35.0-rc.0 (#13103)
- CI: Bump Kubernetes version used for testing to v1.35.0-rc.1 (#13138)
- CI: Drop security scan on 1.10 (#13454)
- CI: Dump resources in scale test (#13232)
- CI: Improve Fake API server (#13183)
- CI: Inmemory APIserver fails for unsupported fieldSelectors (#13306)
- CI: Migrate away from custom GitHub action approval workflow (#13533)
- CI: Revive the debug endpoint for CAPDev in-memory (#13423)
- CI: Use env test 1.35.0 (#13168)
- ClusterCache: Remove stack traces from ClusterCache errors (#13396)
- ClusterClass: Add validation in ClusterClass for CP MachineInfra (#13378)
- ClusterClass: Fix test compute control plane version (#13287)
- ClusterClass: Improve topology diff (#13166)
- ClusterClass: Simplify GetUpgradePlanFromClusterClassVersions (#13276)
- clusterctl: Add cloudscale-ch-cloudscale provider (#13567)