feat: Smooth ClusterProfile/Profile transitions

api/v1beta1/clusterpromotion_types.go

@@ -124,6 +124,16 @@ type ProfileSpec struct {
 	// until all add-ons and applications in the profiles listed as dependencies are deployed.
 	DependsOn []string `json:"dependsOn,omitempty"`
 
+	// TransitionFrom names the Profiles or ClusterProfiles this instance replaces.
+	// For matching target clusters, teardown of the replaced profiles is deferred
+	// until this instance reaches Provisioned. This instance is also permitted to
+	// adopt resources owned by the replaced profiles, ignoring tier restrictions.
+	// Same-kind only: a ClusterProfile names other ClusterProfiles, a Profile names
+	// other Profiles in the same namespace.
+	// +listType=atomic
+	// +optional
+	TransitionFrom []string `json:"transitionFrom,omitempty"`
+
 	// PolicyRefs references all the ConfigMaps/Secrets/Flux Sources containing kubernetes resources
 	// that need to be deployed in the matching managed clusters.
 	// The values contained in those resources can be static or leverage Go templates for dynamic customization.

api/v1beta1/spec.go

@@ -1050,6 +1050,16 @@ type Spec struct {
 	// until all add-ons and applications in the profiles listed as dependencies are deployed.
 	DependsOn []string `json:"dependsOn,omitempty"`
 
+	// TransitionFrom names the Profiles or ClusterProfiles this instance replaces.
+	// For matching target clusters, teardown of the replaced profiles is deferred
+	// until this instance reaches Provisioned. This instance is also permitted to
+	// adopt resources owned by the replaced profiles, ignoring tier restrictions.
+	// Same-kind only: a ClusterProfile names other ClusterProfiles, a Profile names
+	// other Profiles in the same namespace.
+	// +listType=atomic
+	// +optional
+	TransitionFrom []string `json:"transitionFrom,omitempty"`
+
 	// PolicyRefs references all the ConfigMaps/Secrets/Flux Sources containing kubernetes resources
 	// that need to be deployed in the matching managed clusters.
 	// The values contained in those resources can be static or leverage Go templates for dynamic customization.

config/crd/bases/config.projectsveltos.io_clusterprofiles.yaml

@@ -2096,6 +2096,18 @@ spec:
                 format: int32
                 minimum: 1
                 type: integer
+              transitionFrom:
+                description: |-
+                  TransitionFrom names the Profiles or ClusterProfiles this instance replaces.
+                  For matching target clusters, teardown of the replaced profiles is deferred
+                  until this instance reaches Provisioned. This instance is also permitted to
+                  adopt resources owned by the replaced profiles, ignoring tier restrictions.
+                  Same-kind only: a ClusterProfile names other ClusterProfiles, a Profile names
+                  other Profiles in the same namespace.
+                items:
+                  type: string
+                type: array
+                x-kubernetes-list-type: atomic
               validateHealths:
                 description: |-
                   ValidateHealths is a slice of checks to run against the managed cluster

config/crd/bases/config.projectsveltos.io_clusterpromotions.yaml

@@ -1993,6 +1993,18 @@ spec:
                     format: int32
                     minimum: 1
                     type: integer
+                  transitionFrom:
+                    description: |-
+                      TransitionFrom names the Profiles or ClusterProfiles this instance replaces.
+                      For matching target clusters, teardown of the replaced profiles is deferred
+                      until this instance reaches Provisioned. This instance is also permitted to
+                      adopt resources owned by the replaced profiles, ignoring tier restrictions.
+                      Same-kind only: a ClusterProfile names other ClusterProfiles, a Profile names
+                      other Profiles in the same namespace.
+                    items:
+                      type: string
+                    type: array
+                    x-kubernetes-list-type: atomic
                   validateHealths:
                     description: |-
                       ValidateHealths is a slice of checks to run against the managed cluster

config/crd/bases/config.projectsveltos.io_clustersummaries.yaml

@@ -2138,6 +2138,18 @@ spec:
                     format: int32
                     minimum: 1
                     type: integer
+                  transitionFrom:
+                    description: |-
+                      TransitionFrom names the Profiles or ClusterProfiles this instance replaces.
+                      For matching target clusters, teardown of the replaced profiles is deferred
+                      until this instance reaches Provisioned. This instance is also permitted to
+                      adopt resources owned by the replaced profiles, ignoring tier restrictions.
+                      Same-kind only: a ClusterProfile names other ClusterProfiles, a Profile names
+                      other Profiles in the same namespace.
+                    items:
+                      type: string
+                    type: array
+                    x-kubernetes-list-type: atomic
                   validateHealths:
                     description: |-
                       ValidateHealths is a slice of checks to run against the managed cluster

config/crd/bases/config.projectsveltos.io_profiles.yaml

@@ -2096,6 +2096,18 @@ spec:
                 format: int32
                 minimum: 1
                 type: integer
+              transitionFrom:
+                description: |-
+                  TransitionFrom names the Profiles or ClusterProfiles this instance replaces.
+                  For matching target clusters, teardown of the replaced profiles is deferred
+                  until this instance reaches Provisioned. This instance is also permitted to
+                  adopt resources owned by the replaced profiles, ignoring tier restrictions.
+                  Same-kind only: a ClusterProfile names other ClusterProfiles, a Profile names
+                  other Profiles in the same namespace.
+                items:
+                  type: string
+                type: array
+                x-kubernetes-list-type: atomic
               validateHealths:
                 description: |-
                   ValidateHealths is a slice of checks to run against the managed cluster

controllers/clustersummary_controller.go

@@ -430,6 +430,19 @@ func (r *ClusterSummaryReconciler) cleanupBeforeFinalizerRemoval(ctx context.Con
 		return reconcile.Result{Requeue: true, RequeueAfter: deleteRequeueAfter}, nil, true
 	}
 
+	// TransitionFrom support: do not undeploy this profile's resources on this cluster while
+	// a successor profile that currently matches this cluster and names this profile in
+	// TransitionFrom has not yet reached Provisioned. Lets the successor take over shared
+	// resources in place instead of undeploy-then-redeploy.
+	allProvisioned, transitionMsg, err := r.areSuccessorsProvisioned(ctx, clusterSummaryScope, logger)
+	if err != nil {
+		return reconcile.Result{Requeue: true, RequeueAfter: deleteRequeueAfter}, nil, true
+	}
+	clusterSummaryScope.SetDependenciesMessage(&transitionMsg)
+	if !allProvisioned {
+		return reconcile.Result{Requeue: true, RequeueAfter: deleteRequeueAfter}, nil, true
+	}
+
 	// still call undeploy even if cluster is deleted. Sveltos might have deployed resources
 	// in the management cluster and those need to be removed.
 	err = r.undeploy(ctx, clusterSummaryScope, logger)

controllers/clustersummary_transition.go

@@ -0,0 +1,211 @@
+/*
+Copyright 2026. projectsveltos.io. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+*/
+
+package controllers
+
+import (
+	"context"
+	"fmt"
+
+	"github.com/go-logr/logr"
+	corev1 "k8s.io/api/core/v1"
+	apierrors "k8s.io/apimachinery/pkg/api/errors"
+	"k8s.io/apimachinery/pkg/labels"
+	"sigs.k8s.io/controller-runtime/pkg/client"
+
+	configv1beta1 "github.com/projectsveltos/addon-controller/api/v1beta1"
+	"github.com/projectsveltos/addon-controller/lib/clusterops"
+	"github.com/projectsveltos/addon-controller/pkg/scope"
+	libsveltosv1beta1 "github.com/projectsveltos/libsveltos/api/v1beta1"
+	"github.com/projectsveltos/libsveltos/lib/clusterproxy"
+	logs "github.com/projectsveltos/libsveltos/lib/logsettings"
+)
+
+// This file backs TransitionFrom: a ClusterProfile/Profile can name predecessor profile(s) it is
+// replacing on a cluster, so that a plain label swap (moving a cluster from one profile to
+// another that deploys mostly the same resources) does not undeploy-then-redeploy shared
+// resources. See doc/feature_request_clusterprofile_transition.md for the full design.
+
+// transitionSuccessor is a (Cluster)Profile that declares, via TransitionFrom, that it is
+// replacing another (Cluster)Profile.
+type transitionSuccessor struct {
+	kind            string
+	name            string
+	clusterSelector libsveltosv1beta1.Selector
+	clusterRefs     []corev1.ObjectReference
+}
+
+// getTransitionSuccessors returns every (Cluster)Profile that names predecessorName in its
+// TransitionFrom. Same-kind only, matching TransitionFrom's own restriction: a ClusterProfile
+// can only be succeeded by another ClusterProfile, a Profile only by another Profile in the
+// same namespace (a Profile can only ever match clusters in its own namespace to begin with,
+// so scoping this lookup to predecessorNamespace is what keeps the two kinds/namespaces from
+// ever being able to cross).
+func (r *ClusterSummaryReconciler) getTransitionSuccessors(ctx context.Context,
+	predecessorKind, predecessorName, predecessorNamespace string) ([]transitionSuccessor, error) {
+
+	successors := make([]transitionSuccessor, 0)
+
+	if predecessorKind == configv1beta1.ProfileKind {
+		profiles := &configv1beta1.ProfileList{}
+		if err := r.List(ctx, profiles, client.InNamespace(predecessorNamespace)); err != nil {
+			return nil, err
+		}
+		for i := range profiles.Items {
+			p := &profiles.Items[i]
+			if containsName(p.Spec.TransitionFrom, predecessorName) {
+				successors = append(successors, transitionSuccessor{
+					kind: configv1beta1.ProfileKind, name: p.Name,
+					clusterSelector: p.Spec.ClusterSelector, clusterRefs: p.Spec.ClusterRefs,
+				})
+			}
+		}
+		return successors, nil
+	}
+
+	clusterProfiles := &configv1beta1.ClusterProfileList{}
+	if err := r.List(ctx, clusterProfiles); err != nil {
+		return nil, err
+	}
+	for i := range clusterProfiles.Items {
+		cp := &clusterProfiles.Items[i]
+		if containsName(cp.Spec.TransitionFrom, predecessorName) {
+			successors = append(successors, transitionSuccessor{
+				kind: configv1beta1.ClusterProfileKind, name: cp.Name,
+				clusterSelector: cp.Spec.ClusterSelector, clusterRefs: cp.Spec.ClusterRefs,
+			})
+		}
+	}
+	return successors, nil
+}
+
+// isChartTransitioningFrom returns true if claimingHelmManager's owning profile declares, via
+// TransitionFrom, that it is replacing currentHelmManager's owning profile. Mirrors
+// isTransitioningFrom in libsveltos/lib/deployer (used for raw resources), but reads ownership
+// off the two ClusterSummary's own OwnerReferences instead of an annotation on the deployed
+// object, since chartmanager tracks Helm ownership by ClusterSummary rather than by annotation.
+func isChartTransitioningFrom(currentHelmManager, claimingHelmManager *configv1beta1.ClusterSummary) bool {
+	transitionFrom := claimingHelmManager.Spec.ClusterProfileSpec.TransitionFrom
+	if len(transitionFrom) == 0 {
+		return false
+	}
+
+	currentOwner, err := configv1beta1.GetProfileOwnerReference(currentHelmManager)
+	if err != nil || currentOwner == nil {
+		return false
+	}
+	claimingOwner, err := configv1beta1.GetProfileOwnerReference(claimingHelmManager)
+	if err != nil || claimingOwner == nil {
+		return false
+	}
+	// Same-kind only, matching TransitionFrom's own restriction.
+	if currentOwner.Kind != claimingOwner.Kind {
+		return false
+	}
+
+	return containsName(transitionFrom, currentOwner.Name)
+}
+
+func containsName(names []string, target string) bool {
+	for i := range names {
+		if names[i] == target {
+			return true
+		}
+	}
+	return false
+}
+
+// areSuccessorsProvisioned is the mirror of areDependentsRemoved for TransitionFrom: it defers
+// this ClusterSummary's teardown while another (Cluster)Profile that currently matches this
+// cluster names this profile in TransitionFrom and has not reached Provisioned yet, letting the
+// successor take over shared resources in place instead of undeploy-then-redeploy.
+//
+// This cannot key off ClusterSummary existence the way areDependentsRemoved keys off DependsOn:
+// the predecessor and successor profiles are woken by the same label-change event on independent
+// reconcile loops, so the predecessor's teardown can run before the successor has even created
+// its ClusterSummary. At that instant "successor ClusterSummary not found" is indistinguishable
+// from "successor was never going to match this cluster." So this gates on whether the
+// successor's own ClusterSelector/ClusterRefs match this cluster directly instead - that answer
+// is available the instant this reconciles, regardless of whether the successor has reconciled
+// yet.
+//
+// Multiple successors can independently name the same predecessor (fan-in): teardown waits until
+// every currently-matching successor reaches Provisioned, not just one, since each may be
+// responsible for a different subset of the same resources.
+func (r *ClusterSummaryReconciler) areSuccessorsProvisioned(ctx context.Context,
+	clusterSummaryScope *scope.ClusterSummaryScope, logger logr.Logger) (allProvisioned bool, message string, err error) {
+
+	clusterSummary := clusterSummaryScope.ClusterSummary
+
+	profileReference, err := configv1beta1.GetProfileOwnerReference(clusterSummary)
+	if err != nil {
+		logger.V(logs.LogInfo).Info(fmt.Sprintf("failed to get profile owner: %v", err))
+		return false, "", fmt.Errorf("failed to get profile owner: %w", err)
+	}
+	if profileReference == nil {
+		return false, "", fmt.Errorf("profile owner not found")
+	}
+
+	successors, err := r.getTransitionSuccessors(ctx, profileReference.Kind, profileReference.Name,
+		clusterSummary.Namespace)
+	if err != nil {
+		return false, "", err
+	}
+	if len(successors) == 0 {
+		return true, "no transitions pending", nil
+	}
+
+	cluster, err := clusterproxy.GetCluster(ctx, r.Client, clusterSummary.Spec.ClusterNamespace,
+		clusterSummary.Spec.ClusterName, clusterSummary.Spec.ClusterType)
+	if err != nil {
+		if apierrors.IsNotFound(err) {
+			// Cluster is gone. Nothing can still be matching it.
+			return true, "cluster not found", nil
+		}
+		return false, "", err
+	}
+	clusterLabels := labels.Set(cluster.GetLabels())
+
+	for i := range successors {
+		successor := successors[i]
+		if !selectorMatchesCluster(successor.clusterSelector, successor.clusterRefs, clusterSummary, clusterLabels) {
+			// This successor does not (or no longer) match this cluster: it is not part of a
+			// transition on this cluster, whether or not it is still Provisioned elsewhere.
+			continue
+		}
+
+		successorClusterSummary, err := clusterops.GetClusterSummary(ctx, r.Client, successor.kind, successor.name,
+			clusterSummary.Spec.ClusterNamespace, clusterSummary.Spec.ClusterName, clusterSummary.Spec.ClusterType)
+		if err != nil {
+			if apierrors.IsNotFound(err) {
+				msg := fmt.Sprintf("%s %s is replacing this profile on this cluster but has not deployed yet",
+					successor.kind, successor.name)
+				logger.V(logs.LogInfo).Info(msg)
+				return false, msg, nil
+			}
+			return false, "", err
+		}
+
+		if !isCluterSummaryProvisioned(successorClusterSummary) {
+			msg := fmt.Sprintf("%s %s is replacing this profile on this cluster but is not fully deployed yet",
+				successor.kind, successor.name)
+			logger.V(logs.LogInfo).Info(msg)
+			return false, msg, nil
+		}
+	}
+
+	return true, "all transitioning successors are provisioned", nil
+}