fix: deletion of zero sized nodepools (#2049)

If a nodepool was scheduled that had 0 nodes, the actual deletion of the
nodepool from the current state would be skipped resulting in an
infinite loop of scheduling its deletion. This PR adds an explicit
deletion from the current state even if the nodepool is zero-sized if a
deletion for it has been scheduled.

<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit

* **Bug Fixes**
* Ensure nodepool entries are removed from cluster configuration and the
updated Kubernetes state is persisted when removing nodes and no
masters/workers remain.
* **Chores**
* Bumped container image tags for several components (including testing
framework and core services) to a new release.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->

---------

Co-authored-by: CI/CD pipeline <CI/CD-pipeline@users.noreply.github.com>

Author: Despire

manifests/claudie/kustomization.yaml

@@ -56,16 +56,16 @@ apiVersion: kustomize.config.k8s.io/v1beta1
 kind: Kustomization
 images:
 - name: ghcr.io/berops/claudie/ansibler
-  newTag: fe5a65d-4071
+  newTag: c71fbb0-4077
 - name: ghcr.io/berops/claudie/autoscaler-adapter
-  newTag: fe5a65d-4071
+  newTag: c71fbb0-4077
 - name: ghcr.io/berops/claudie/claudie-operator
-  newTag: fe5a65d-4071
+  newTag: c71fbb0-4077
 - name: ghcr.io/berops/claudie/kube-eleven
-  newTag: fe5a65d-4071
+  newTag: c71fbb0-4077
 - name: ghcr.io/berops/claudie/kuber
-  newTag: fe5a65d-4071
+  newTag: c71fbb0-4077
 - name: ghcr.io/berops/claudie/manager
-  newTag: fe5a65d-4071
+  newTag: c71fbb0-4077
 - name: ghcr.io/berops/claudie/terraformer
-  newTag: fe5a65d-4071
+  newTag: c71fbb0-4077

manifests/testing-framework/kustomization.yaml

@@ -89,4 +89,4 @@ secretGenerator:
 
 images:
 - name: ghcr.io/berops/claudie/testing-framework
-  newTag: fe5a65d-4071
+  newTag: c71fbb0-4077

services/kuber/internal/worker/service/task_delete_nodes.go

@@ -82,6 +82,18 @@ func DeleteNodes(logger zerolog.Logger, tracker Tracker) {
 	}
 
 	if len(master) == 0 && len(worker) == 0 {
+		if d.KDeleteNodes.WithNodePool {
+			logger.
+				Info().
+				Msgf("Removing nodepool %q with 0 nodes", d.KDeleteNodes.Nodepool)
+
+			k8s.ClusterInfo.NodePools = nodepools.DeleteByName(k8s.ClusterInfo.NodePools, d.KDeleteNodes.Nodepool)
+
+			update := tracker.Result.Update()
+			update.Kubernetes(k8s)
+			update.Commit()
+		}
+
 		return
 	}