feat: determine cluster ID from cluster name (label) if in LKE mode

.github/workflows/image-build-push.yml

@@ -0,0 +1,31 @@
+name: Image
+on:
+  push:
+    branches:
+      - "main"
+    tags:
+      - "*"
+
+permissions:
+  contents: write
+  actions: read
+
+jobs:
+  release:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v6
+        with:
+          fetch-depth: 0
+
+      - name: Login to Docker Hub
+        uses: docker/login-action@v3
+        with:
+          username: ${{ secrets.DOCKER_USERNAME }}
+          password: ${{ secrets.DOCKER_PASSWORD }}
+
+      - name: Install devbox
+        uses: jetify-com/devbox-install-action@v0.14.0
+
+      - name: Build and push image
+        run: devbox run make image

.github/workflows/release.yml

@@ -16,18 +16,6 @@ jobs:
         with:
           fetch-depth: 0
 
-      - name: Login to Docker Hub
-        uses: docker/login-action@v3
-        with:
-          username: ${{ secrets.DOCKER_USERNAME }}
-          password: ${{ secrets.DOCKER_PASSWORD }}
-
-      - name: Install devbox
-        uses: jetify-com/devbox-install-action@v0.14.0
-
-      - name: Build and push image
-        run: devbox run make image
-
       - name: Create Release Artifacts
         run: devbox run make release
         env:

Makefile

@@ -143,7 +143,6 @@ helm-install: ## install karpenter onto an existing cluster (requires k8s contex
 	@helm upgrade --install --namespace karpenter --create-namespace karpenter charts/karpenter \
 		--set controller.image.repository=$(KO_DOCKER_REPO) \
 		--set region=${LINODE_REGION} \
-		--set settings.clusterID=${CLUSTER_ID} \
 		--set settings.clusterName=${CLUSTER_NAME} \
 		--set apiToken=${LINODE_TOKEN}
 

charts/karpenter/templates/karpl-credentials.yaml

@@ -7,6 +7,5 @@ stringData:
   LINODE_TOKEN: {{ required "The API token needs to be set" .Values.apiToken }}
   LINODE_URL: {{ .Values.apiURL }}
   LINODE_API_VERSION: {{ .Values.apiVersion }}
-  CLUSTER_ID: {{ required "The LKE cluster ID needs to be set" .Values.settings.clusterID | quote }}
   # TODO: region discovery
   CLUSTER_REGION: {{ required "The cluster region currently needs to be set" .Values.region }}

charts/karpenter/values.yaml

@@ -166,8 +166,6 @@ settings:
   clusterCABundle: ""
   # -- Cluster name.
   clusterName: ""
-  # -- LKE cluster ID (required when mode is 'lke').
-  clusterID: ""
   # -- Cluster endpoint.
   clusterEndpoint: ""
   # -- Operating mode: 'lke' for LKE NodePool provisioning, 'instance' for direct Linode instances.

cmd/controller/main.go

@@ -37,7 +37,11 @@ import (
 func main() {
 	ctx1, op1 := coreoperator.NewOperator()
 	linodeClientConfig := validateEnvironment(ctx1)
-	ctx, op := operator.NewOperator(ctx1, op1, linodeClientConfig)
+	ctx, op, err := operator.NewOperator(ctx1, op1, linodeClientConfig)
+	if err != nil {
+		log.FromContext(ctx1).Error(err, "Failed to create Linode operator")
+		os.Exit(1)
+	}
 
 	linodeCloudProvider := cloudprovider.New(
 		op.InstanceTypesProvider,

hack/boilerplate.sh

@@ -2,7 +2,7 @@
 set -eu -o pipefail
 
 for i in $(
-  find ./tools ./cmd ./pkg ./test ./hack -name "*.go"
+  find ./cmd ./pkg ./hack -name "*.go"
 ); do
   if ! grep -q "Apache License" $i; then
     cat hack/boilerplate.go.txt $i >$i.new && mv $i.new $i

pkg/fake/linodeapi.go

@@ -671,3 +671,8 @@ func (l *LinodeClient) GetLKENodePoolNode(_ context.Context, clusterID int, node
 	}
 	return *node, err
 }
+
+func (l *LinodeClient) ListLKEClusters(_ context.Context, _ *linodego.ListOptions) ([]linodego.LKECluster, error) {
+	// For simplicity, return a canned response
+	return []linodego.LKECluster{{ID: DefaultClusterID}}, nil
+}

pkg/linode/sdk.go

@@ -35,7 +35,10 @@ type LinodeAPI interface {
 	CreateTag(ctx context.Context, opts linodego.TagCreateOptions) (*linodego.Tag, error)
 	CreateInstance(ctx context.Context, opts linodego.InstanceCreateOptions) (*linodego.Instance, error)
 
+	// LKE Cluster methods for LKE cluster management
+	ListLKEClusters(ctx context.Context, opts *linodego.ListOptions) ([]linodego.LKECluster, error)
 	// NodePool methods for LKE cluster management
+
 	CreateLKENodePool(ctx context.Context, clusterID int, opts linodego.LKENodePoolCreateOptions) (*linodego.LKENodePool, error)
 	ListLKENodePools(ctx context.Context, clusterID int, opts *linodego.ListOptions) ([]linodego.LKENodePool, error)
 	GetLKENodePool(ctx context.Context, clusterID, poolID int) (*linodego.LKENodePool, error)

pkg/operator/operator.go

@@ -19,6 +19,7 @@ import (
 	"fmt"
 	"net"
 
+	"github.com/linode/linodego"
 	"github.com/patrickmn/go-cache"
 	"github.com/samber/lo"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
@@ -32,6 +33,7 @@ import (
 	"github.com/linode/karpenter-provider-linode/pkg/providers/instance"
 	"github.com/linode/karpenter-provider-linode/pkg/providers/instancetype"
 	"github.com/linode/karpenter-provider-linode/pkg/providers/lke"
+	"github.com/linode/karpenter-provider-linode/pkg/utils"
 )
 
 func init() {
@@ -47,7 +49,7 @@ type Operator struct {
 	LinodeClient              sdk.LinodeAPI
 }
 
-func NewOperator(ctx context.Context, operator *operator.Operator, linodeClientConfig sdk.ClientConfig) (context.Context, *Operator) {
+func NewOperator(ctx context.Context, operator *operator.Operator, linodeClientConfig sdk.ClientConfig) (context.Context, *Operator, error) {
 	linodeClient := lo.Must(sdk.CreateLinodeClient(linodeClientConfig))
 	unavailableOfferingsCache := linodecache.NewUnavailableOfferings()
 	kubeDNSIP, err := KubeDNSIP(ctx, operator.KubernetesInterface)
@@ -79,8 +81,24 @@ func NewOperator(ctx context.Context, operator *operator.Operator, linodeClientC
 	switch opts.Mode {
 	case "lke":
 		log.FromContext(ctx).Info("initializing in LKE mode")
+		listFilter := utils.Filter{Label: opts.ClusterName}
+		filter, err := listFilter.String()
+		if err != nil {
+			return nil, nil, err
+		}
+		// Get cluster ID from the cluster name (label)
+		clusterList, err := linodeClient.ListLKEClusters(ctx, &linodego.ListOptions{
+			Filter: filter,
+		})
+		if err != nil {
+			return nil, nil, err
+		}
+		if len(clusterList) != 1 {
+			return nil, nil, fmt.Errorf("could not determine LKE cluster with name: %s", opts.ClusterName)
+		}
+
 		nodeProvider = lke.NewDefaultProvider(
-			opts.ClusterID,
+			clusterList[0].ID,
 			opts.ClusterRegion,
 			operator.EventRecorder,
 			linodeClient,
@@ -105,7 +123,7 @@ func NewOperator(ctx context.Context, operator *operator.Operator, linodeClientC
 		InstanceTypesProvider:     instanceTypeProvider,
 		NodeProvider:              nodeProvider,
 		LinodeClient:              linodeClient,
-	}
+	}, nil
 }
 
 func KubeDNSIP(ctx context.Context, kubernetesInterface kubernetes.Interface) (net.IP, error) {

pkg/operator/options/options.go

@@ -35,7 +35,6 @@ type optionsKey struct{}
 
 type Options struct {
 	ClusterName             string
-	ClusterID               int
 	ClusterRegion           string
 	ClusterEndpoint         string
 	VMMemoryOverheadPercent float64
@@ -45,7 +44,6 @@ type Options struct {
 
 func (o *Options) AddFlags(fs *coreoptions.FlagSet) {
 	fs.StringVar(&o.ClusterName, "cluster-name", env.WithDefaultString("CLUSTER_NAME", ""), "[REQUIRED] The kubernetes cluster name for resource discovery.")
-	fs.IntVar(&o.ClusterID, "cluster-id", env.WithDefaultInt("CLUSTER_ID", 0), "The LKE cluster ID for LKE node pool operations. Required when mode is 'lke'.")
 	fs.StringVar(&o.ClusterRegion, "cluster-region", env.WithDefaultString("CLUSTER_REGION", "us-east"), "The region the kubernetes cluster is deployed in.")
 	fs.StringVar(&o.ClusterEndpoint, "cluster-endpoint", env.WithDefaultString("CLUSTER_ENDPOINT", ""), "The external kubernetes cluster endpoint for new nodes to connect with. If not specified, will discover the cluster endpoint using DescribeCluster API.")
 	fs.Float64Var(&o.VMMemoryOverheadPercent, "vm-memory-overhead-percent", utils.WithDefaultFloat64("VM_MEMORY_OVERHEAD_PERCENT", 0.075), "The VM memory overhead as a percent that will be subtracted from the total memory for all instance types when cached information is unavailable.")

pkg/operator/options/options_validation.go

@@ -55,9 +55,6 @@ func (o *Options) validateRequiredFields() error {
 	if o.ClusterName == "" {
 		return fmt.Errorf("missing field, cluster-name")
 	}
-	if o.Mode == "lke" && o.ClusterID <= 0 {
-		return fmt.Errorf("missing or invalid field, cluster-id must be a positive integer when mode is 'lke'")
-	}
 	return nil
 }
 

pkg/operator/options/suite_test.go

@@ -57,14 +57,12 @@ var _ = Describe("Options", func() {
 		opts.AddFlags(fs)
 		err := opts.Parse(fs,
 			"--cluster-name", "env-cluster",
-			"--cluster-id", "12345",
 			"--cluster-endpoint", "https://env-cluster",
 			"--cluster-region", "us-west",
 			"--vm-memory-overhead-percent", "0.1")
 		Expect(err).ToNot(HaveOccurred())
 		expectOptionsEqual(opts, test.Options(test.OptionsFields{
 			ClusterName:             lo.ToPtr("env-cluster"),
-			ClusterID:               lo.ToPtr(12345),
 			ClusterEndpoint:         lo.ToPtr("https://env-cluster"),
 			ClusterRegion:           lo.ToPtr("us-west"),
 			VMMemoryOverheadPercent: lo.ToPtr[float64](0.1),
@@ -84,7 +82,6 @@ var _ = Describe("Options", func() {
 		Expect(err).ToNot(HaveOccurred())
 		expectOptionsEqual(opts, test.Options(test.OptionsFields{
 			ClusterName:             lo.ToPtr("env-cluster"),
-			ClusterID:               lo.ToPtr(12345),
 			ClusterEndpoint:         lo.ToPtr("https://env-cluster"),
 			ClusterRegion:           lo.ToPtr("us-west"),
 			VMMemoryOverheadPercent: lo.ToPtr[float64](0.1),
@@ -100,11 +97,11 @@ var _ = Describe("Options", func() {
 			Expect(err).To(HaveOccurred())
 		})
 		It("should fail when clusterEndpoint is invalid (not absolute)", func() {
-			err := opts.Parse(fs, "--cluster-name", "test-cluster", "--cluster-id", "12345", "--cluster-endpoint", "00000000000000000000000.test.us-west.linode.com")
+			err := opts.Parse(fs, "--cluster-name", "test-cluster", "--cluster-endpoint", "00000000000000000000000.test.us-west.linode.com")
 			Expect(err).To(HaveOccurred())
 		})
 		It("should fail when vmMemoryOverheadPercent is negative", func() {
-			err := opts.Parse(fs, "--cluster-name", "test-cluster", "--cluster-id", "12345", "--vm-memory-overhead-percent", "-0.01")
+			err := opts.Parse(fs, "--cluster-name", "test-cluster", "--vm-memory-overhead-percent", "-0.01")
 			Expect(err).To(HaveOccurred())
 		})
 	})
@@ -113,7 +110,6 @@ var _ = Describe("Options", func() {
 func expectOptionsEqual(optsA *options.Options, optsB *options.Options) {
 	GinkgoHelper()
 	Expect(optsA.ClusterName).To(Equal(optsB.ClusterName))
-	Expect(optsA.ClusterID).To(Equal(optsB.ClusterID))
 	Expect(optsA.ClusterEndpoint).To(Equal(optsB.ClusterEndpoint))
 	Expect(optsA.VMMemoryOverheadPercent).To(Equal(optsB.VMMemoryOverheadPercent))
 	Expect(optsA.ClusterRegion).To(Equal(optsB.ClusterRegion))

pkg/test/options.go

@@ -26,7 +26,6 @@ import (
 
 type OptionsFields struct {
 	ClusterName             *string
-	ClusterID               *int
 	ClusterEndpoint         *string
 	ClusterRegion           *string
 	VMMemoryOverheadPercent *float64
@@ -42,7 +41,6 @@ func Options(overrides ...OptionsFields) *options.Options {
 	}
 	return &options.Options{
 		ClusterName:             lo.FromPtrOr(opts.ClusterName, "123456789012"),
-		ClusterID:               lo.FromPtrOr(opts.ClusterID, fake.DefaultClusterID),
 		ClusterEndpoint:         lo.FromPtrOr(opts.ClusterEndpoint, "https://test-cluster"),
 		ClusterRegion:           lo.FromPtrOr(opts.ClusterRegion, fake.DefaultRegion),
 		VMMemoryOverheadPercent: lo.FromPtrOr(opts.VMMemoryOverheadPercent, 0.075),