Tasks for v1.9 release cycle

[Sunnatillo]

Please see the corresponding sections of the role-handbooks for documentation of individual tasks.

Tasks

Notes:

• Weeks are only specified to give some orientation.
• The following is based on the v1.9 release cycle. Modify according to the tracked release cycle.

Week 1:

• [Release Lead] Finalize release schedule and team
  • 🌱 Add v1.9 release team members to timeline doc #11097
• [Release Lead] Add/remove release team members
  • 🌱 Swap in new 1.9 RT members in OWNERS #11096
• [Release Lead] Prepare main branch for development of the new release
  • PR: 🌱 Prepare main branch for v1.9 development #11059
• [Communications Manager] Add docs to collect release notes for users and migration notes for provider implementers
  • 📖 Add more documentation about the KCP pre-terminate hook #11153
• [Communications Manager] Update supported versions
  • 📖 Update CAPI support and guarantees for v1.9 #11163

Week 1 to 4:

• [Release Lead] [Track] Remove previously deprecated code
  • 🌱 Remove previously deprecated --metrics-bind-addr flag #11140

Week 4:

• [Release Lead] Cut the v1.7.6 & v1.8.3 releases
  • 🚀 Release v1.7.6 #11158
  • 🚀 Release v1.8.3 #11160

Week 8:

• [Release Lead] Cut the v1.7.7 & v1.8.4 releases
  • 🚀 Release v1.8.4 #11269
  • 🚀 Release v1.7.7 #11268

Week 11 to 12:

• [Release Lead] [Track] Bump dependencies
• 🌱 Bump Go to v1.22.8 version #11357

Week 12:

• [Release Lead] Cut the v1.9.0-beta.0 release
  • 🚀 Release v1.9.0-beta.0 #11371
• [Release Lead] Cut the v1.7.8 & v1.8.5 releases
  • 🚀 Release v1.8.5 #11373
  • 🚀 Release v1.7.8 #11372
• [Release Lead] Create a new GitHub milestone for the next release
• [Communications Manager] Communicate beta to providers

Week 13:

• KubeCon idle week

Week 14:

• [Release Lead] Cut the v1.9.0-beta.1 release
  • 🚀 Release v1.9.0-beta.1 #11441
• [Release Lead] Set a tentative release date for the next minor release
• [Release Lead] Assemble next release team
• [Release Lead] Select release lead for the next release cycle

Week 15:

• [Release Lead] Cut the v1.9.0-rc.0 release
  • 🚀 Release v1.9.0-rc.0 #11469
• [Release Lead] Update milestone applier and GitHub Actions
  • Update milestone applier for CAPI 1.9 and 1.10 kubernetes/test-infra#33861
• [CI Manager] Setup jobs and dashboards for the release-1.9 release branch
  • 🌱 Add jobs for CAPI release 1.9 and remove 1.6 jobs kubernetes/test-infra#33816
• [Communications Manager] Ensure the book for the new release is available
  • 📖 add release 1.8 book link #11486

Week 16:

• [Release Lead] Cut the v1.9.0-rc.1 release
  • 🚀 Release v1.9.0-rc.1 #11519

Week 17:

• [Release Lead] Cut the v1.9.0 release
  • 🚀 Release v1.9.0 #11554
• [Release Lead] Cut the v1.7.9 & v1.8.6 releases
  • 🚀 Release v1.7.9 #11555
  • 🚀 Release v1.8.6 #11556
• [Release Lead] Organize release retrospective
• [Communications Manager] Change production branch in Netlify to the new release branch
• [Communications Manager] Update clusterctl links in the quickstart
  • 📖 Update clusterctl version to v1.9.x in quick start #11559

Continuously:

• [Release lead] Maintain the GitHub release milestone
• [Release lead] Bump the Go version
• [Communications Manager] Communicate key dates to the community
• [Communications Manager] Improve release process documentation
• [Communications Manager] Maintain and improve user facing documentation about releases, release policy and release calendar
• [CI Manager] Monitor CI signal
• [CI Manager] Reduce the amount of flaky tests

If and when necessary:

• [Release Lead] [Track] Bump the Cluster API apiVersion
• [Release Lead] [Track] Bump the Kubernetes version
• [Release Lead] Track Release and Improvement tasks

/priority critical-urgent
/kind feature

Reminder!
Add to release notes:

Highlights

• Bumped to controller-runtime v0.19, k8s.io/* v0.31, controller-gen v0.16
• Improvements to status fields: (Tracking issue for Improving status in CAPI resources - Phase 1 #11105)
  • Proposal: Improving status in CAPI resources
  • Just a few highlights:
    • Overhaul of our contract documentation: Provider contracts
    • Implemented v1beta2 conditions for core Cluster API objects
    • Huge improvements to Deleting conditions of Cluster, KCP, MD, MS and Machines
    • Added Available conditions to Cluster, KCP and MD
    • Added ControlPlaneInitialized, ControlPlaneAvailable, WorkersAvailable, RemoteConnectionProbe, conditions to Cluster
    • Added UpToDate conditions to KCP and MD Machines
    • Added ScalingUp, ScalingDown, Remediating, MachinesReady, MachinesUpToDate conditions to Cluster, KCP, MD and MS
    • Significantly improved Available, EtcdClusterHealthy and ControlPlaneComponentsHealthy conditions on KCP
    • Added Paused conditions
    • Improved counter fields for Cluster, KCP, MD, MS
    • Added utils and extended patch helper for v1beta2 conditions
• ClusterCache supersedes the now deprecated ClusterCacheTracker (Improve ClusterCacheTracker to address fundamental limitations #11272)
• ClusterClass: Support oneOf/anyOf/allOf/not in variable schemas (✨ Add support for oneOf/anyOf/allOf/not ClusterClass variable schema constructs #10637)
• Improved logging in predicates (🌱 Log controller name in predicates #10959 ⚠️ Fix object logging in predicates #11239 ⚠️ Fix object logging in ResourceHasFilterLabel, ResourceNotPaused & ResourceNotPausedAndHasFilterLabel predicates #11188)
• KubeadmControlPlane: Implement pre-terminate hook for clean Machine Deletion (🐛 KCP: remove etcd member in pre-terminate hook #11137)
• KubeadmControlPlane: NamingStrategy for Machines (🌱 Add NamingStrategy to KubeadmControlPlane #11123)
• MachineSetPreflightChecks feature got promoted to beta (✨ MachineSet: Promote MachineSetPreflightChecks to Beta #11228)
• MachineSet/MachineDeployment: Foreground deletion for MachineDeployments and MachineSets (🌱 Foreground deletion for MachineDeployments and MachineSets #11174)
• Machine: Improvements to Node drain:
  • Improve Node drain observability (✨ Refactor node drain #11074 📖 Add Node drain documentation #11121): documentation: Machine deletion process
  • MachineDrainRules Configurable Machine drain behavior (Configurable Machine drain behavior #11240): Proposal: MachineDrainRules
  • Stop waiting for detachment of volumes belonging to Pods ignored during drain (⚠️ Machine: ignore attached Volumes referred by pods ignored during drain #11246 🌱 Add feature gate to consider VolumeAttachments when waiting for volume detach #11386)
• As usual, significantly improved e2e test framework & coverage (e.g. ✨ Detect panics in e2e tests #11066 🌱 Export scale e2e test to make it reusable by providers #11333 🌱 test: add coverage to find issues running on the next kubernetes release as management cluster #10984 🌱 test: improve autoscale tests for to/from zero and running autoscaler in bootstrap cluster #11082 🌱 test: use cluster without worker definition for kcp only test #11212 🌱 Improve node drain e2e test #11127 🌱 Extend Node drain e2e test to cover MachineDrainRules #11362)
• New providers in clusterctl: Canonical (🌱 Add Canonical Kubernetes providers #11337), Nutanix (🌱 clusterctl: add nutanix ipam & runtime extensions providers #11135), Vultr (🌱 Add vultr infrastructure provider #11091)

Deprecation and Removals Warning

• API: Deprecated FailureMessage and FailureReason fields (🌱 api: deprecate FailureMessage and FailureReason #11317)
• The ClusterCacheTracker component has been deprecated (🌱 Deprecate ClusterCacheTracker #11312 🌱 Deprecate CCT ErrClusterLocked #11340), please use the new ClusterCache instead.
  • For more context and examples for how to use it, see PR: Introduce new ClusterCache and the corresponding issue
  • Please note:
    • The DisabledFor option (previously ClientUncachedObjects) is not defaulted to &corev1.ConfigMap & &corev1.Secret anymore, thus it’s now necessary to explicitly set DisabledFor to avoid caching ConfigMaps and Secrets.
    • SecretClient and UserAgent are now mandatory options, please take a look at the corresponding godoc.
  • The --clustercachetracker-concurrency, --clustercachetracker-client-qps and --clustercachetracker-client-burst flags on all core CAPI controllers have been renamed to --clustercache-concurrency, --clustercache-client-qps and --clustercache-client-burst
• The deprecated --metrics-bind-addr flag has been removed (🌱 Remove previously deprecated --metrics-bind-addr flag #11140)
• Deprecated obsolete errors package (🌱 Deprecate obsolete errors pkg #10798)
• Deprecated ClusterCreateInfraReady and ClusterUnpausedAndInfrastructureReady predicates (🌱 Deprecate old pausing predicates #11300)

[sbueringer]

/triage accepted

(@Sunnatillo just fyi, I removed the spam)

[sbueringer]

@Sunnatillo fyi, I updated the issue description according to #11088 (mostly to fix the links)

[Sunnatillo]

I have fixed the week numbers to align with https://github.com/kubernetes-sigs/cluster-api/blob/main/docs/release/releases/release-1.9.md