Tracking issue for Improving status in CAPI resources

[fabriziopandini]

This is a tracking issue for activities related to #10897

Enablement

Activities that are required to unblock implementation

• @fabriziopandini Implement conditions utils ✨ Implement utils for v1beta2 conditions #10997
• @fabriziopandini Add support for metav1 conditions in patch helper 🌱 Add support for v1beta2 conditions to patch helper #11150
  • Audit usage of forceOverwrite
    • rif 🌱 Add support for v1beta2 conditions to patch helper #11150 (comment)
• @sbueringer Improve Cluster Cache Tracker to allow tracking of connection problems ✨ Introduce new ClusterCache #11247
• @chrischdi MD foreground deletion 🌱 Foreground deletion for MachineDeployments and MachineSets #11174
• @sbueringer Handle finalizers early in Reconciles 🌱 Handle finalizers early in Reconciles #11286

Documentation updates

• @fabriziopandini add superseded by notice to old condition proposal PR - no backport required
• @fabriziopandini update provider contracts with notices about future changes & clarify requirements about conditions in provider contracts. Note, no backport is planned
  • @fabriziopandini 📖 Drop k8s.gcr.io deprecation notice #11190
  • @fabriziopandini 📖 Drop old version of migration docs #11191
  • @fabriziopandini 📖 Group API reference docs #11194
  • @fabriziopandini 📖 Refactor provider's implementers guide #11195
  • @fabriziopandini 📖 Refactor core developer guide #11208
  • @fabriziopandini 📖 Refactor InfraCluster contract #11213
  • @fabriziopandini 📖 Refactor InfraMachine contract #11223
  • @fabriziopandini 📖 Refactor the BootstrapConfig contract #11235
  • @fabriziopandini 📖 Refactor ControlPlane contract #11285
  • machine pool contract

Implementation Phase1

Focus on phase 1 of the implementation, where new field are added but not yet used by controllers

• API modeling
  • @fabriziopandini ✨ Add v1beta2 structs to object status #11234
    • Add to condition fields godoc "Known condition types are Foo, Bar and Baz."
      • rif. 📖 Proposal: Improving status in CAPI resources #10897 (comment)
    • Fine tuning of godocs for API trypes (make them kubectl explain compliant)
      • rif 📖 Proposal: Improving status in CAPI resources #10897 (comment)
      • might be relevant in the context of Umbrella issue: API changes #10852
  • Deprecate FailureReason, Message (others?) @chrischdi
    • 🌱 api: deprecate FailureMessage and FailureReason #11317
• Machine controller
  • @fabriziopandini "Core" conditions: bootstrapReady, InfraReady, NodeHealthy, NodeReady, Paused ✨ Update machine with v1beta2 status #11276
    • Available, ready + ReadinessGates
  • Up to date condition
    • @fabriziopandini In place propagation of minReadySeconds down to machine (from KCP. MD, MP)
    • @fabriziopandini In place propagation of UpToDate condition from owning controllers (from KCP. MD, MP)
      • Improve UpToDate condition to surface if a rolling upgrade and it was triggered by an infraMachineTemplate change or a kube version change, etc (KCP. MD, MP)
  • @sbueringer Deleting: 🌱 Implement Machine Deleting condition #11291
  • @sbueringer Make node conditions to consider connection state 🌱 Implement Machine Deleting condition #11291
• @fabriziopandini MachineHealthCheck controller ✨ Update MHC with v1Beta2 status #11290
  • HealthCheckSucceeded and OwnerRemediated
  • MHC owns conditions
• MachineSet controller
  • "Core" conditions: ScalingUp, ScalingDown, Paused @chrischdi
    • ✨ Update machineset with v1beta2 status test #11278
  • MachinesReady, MachinesUpToDate (depends on Machine controller work)
  • Conditions set at machine level
    • OwnerRemediated
  • Remediating
  • Deleting
  • Counter fields
  • Improve ScaleUp/ScaleDown/Remediating if blocked by preflight check
• MachineDeployment controller
  • "Core" conditions: Available, ScalingUp, ScalingDown, Paused
  • MachinesReady, MachinesUpToDate (depends on Machine controller work)
  • Remediating
  • Deleting
  • Counter fields
• @fabriziopandini KubeadmControlPlane controller
  • "Core" conditions: ScalingUp, ScalingDown, Paused, CertificatesAvailable
  • Available
  • MachinesReady, MachinesUpToDate (depends on Machine controller work)
  • Remediating
  • Deleting
  • Counter fields
  • Enforce readiness gates at machine level
  • Conditions set at machine level + ReadinessGates
    • @fabriziopandini remediation ✨ Add v1beta2 OwnerRemediated condition to KCP #11297
    • @fabriziopandini etcd & control plane components ✨ Add v1beta2 Etcd and ControlPlaneComponents conditions to KCP #11302
    • Make etcd & control plane conditions to consider connection state
• Cluster controller (Topology?)
  • "Core" conditions: ScalingUp, ScalingDown, Initialized, InfrastructureReady, Paused, TopologyReconciled
  • @sbueringer RemoteConnectionProbe ✨ Introduce new ClusterCache #11247
  • MachinesReady, MachinesUpToDate (depends on Machine controller work)
  • Available, WorkersAvailable, ControlPlaneAvailable + Availability gates
  • Remediating
  • Deleting
  • Counter fields
• Topology controller
  • Let’s surface Paused in TopologyReconciled condition (let's do this after implementing TopologyReconciled itself)
• ClusterResourceSet controller
• ClusterClass controller
• DockerCluster and DockerMachine controllers
• MachinePool conditions
  • ...
• @fabriziopandini Review clusterctl describe cluster [WIP] ✨ Add support for v1beta2 conditions to clusterctl describe #11277

Follow up / before release

Activities that could be deferred to after we get a first version of this proposal implemented

• Prune fields not yet implemented before release
• Add support for negative polarity conditions to ReadinessGates and AvailabilityGates
  • rif 📖 Proposal: Improving status in CAPI resources #10897 (review)
• Extending CC to allow the definition of AvailabilityGates & ReadinessGates (and Cluster topology)
  • rif 📖 Proposal: Improving status in CAPI resources #10897 (comment)
  • rif ✨ Add v1beta2 structs to object status #11234 (comment)
• Allow customisation of remote-conditions-grace-period at cluster level
  • rif 📖 Proposal: Improving status in CAPI resources #10897 (comment)
• Consider if to allow users to add arbitrary ReadinessGates to KCP machines
• Rename test types to align to API types (v1Beta2 in the middle)
• Check if NodeNetworkUnavailable must be added to MachineNodeHealthyCondition
• ExtensionConfig
  • Add v1beta2 struct
  • Paused
• DockerMachinePool
  • Add v1beta2 struct
  • Paused
• ClusterResourceSetBinding
  • Add v1beta2 struct
  • Paused
• Check initializations of the status structs to catch cases like this

Implementation Phase2

Initial notes for phase 2 of the implementation, where new field are added but not yet used by controllers

• ...

• API modeling

  • Introduce deprecated struct, move v1beta2 top level
  • Rename condition interfaces and method implementations (current--> deprecated, v1beta2--> current)
  • Rename condition consts (current--> deprecated, v1beta2--> current)
  • Move conditions fields as the first field in the struct
    • rif. 📖 Proposal: Improving status in CAPI resources #10897 (comment)
  • Fine tuning of print columns
    • rif 📖 Proposal: Improving status in CAPI resources #10897 (review)
    • rif 📖 Proposal: Improving status in CAPI resources #10897 (comment)

• Change contracts

  • ...

• Use new fields

  • MachineHealthCheck controller
    • Change test for ControlPlaneInitializedCondition in reconcile
    • Get MachineHealthCheckSucceededCondition in patchUnhealthyTargets
    • Drop failureReson/failureMessage test from needsRemediation
    • Change test for ControlPlaneInitializedCondition, InfrastructureReadyCondition in needsRemediation (few places)
  • KCP controller
    • Cahnge test for MachineEtcdMemberHealthyCondition in canSafelyRemoveEtcdMember + add a test for ClusterHealthy if not already there

• Condition utils

  • Rename packages for condition utils (conditions-->deprecated conditions, v1beta2conditions --> conditions)

• Clusterctl

  • clusterctl describe (drop v1beta2 flag, introduce deprecated flag, switch default behaviour in the implementation)

Implementation Phase3

• ...

Backlog / unprioritized

• Scale testing with CAPD (check for condition flakiness)

[sbueringer]

Feel free to assign me to "Improve Cluster Cache Tracker to allow tracking of connection problems"