feat: Add ability to select instance by EBS Maximum Bandwidth (#5925)

aws · May 5, 2024 · b17e0eb · b17e0eb
1 parent 39057a5
commit b17e0eb
Show file tree

Hide file tree

Showing 13 changed files with 958 additions and 6 deletions.
diff --git a/designs/v1beta1-api.md b/designs/v1beta1-api.md
@@ -346,6 +346,7 @@ status:
 8. `karpenter.k8s.aws/instance-cpu`
 9. `karpenter.k8s.aws/instance-cpu-manufacturer`
 10. `karpenter.k8s.aws/instance-memory`
+11. `karpenter.k8s.aws/instance-ebs-bandwidth`
 11. `karpenter.k8s.aws/instance-network-bandwidth`
 12. `karpenter.k8s.aws/instance-gpu-name`
 13. `karpenter.k8s.aws/instance-gpu-manufacturer`

diff --git a/hack/code/instancetype_testdata_gen/main.go b/hack/code/instancetype_testdata_gen/main.go
@@ -130,6 +130,23 @@ func getInstanceTypeInfo(info *ec2.InstanceTypeInfo) string {
 	fmt.Fprintf(src, "SizeInMiB: aws.Int64(%d),\n", lo.FromPtr(info.MemoryInfo.SizeInMiB))
 	fmt.Fprintf(src, "},\n")
 
+	if info.EbsInfo != nil {
+		fmt.Fprintf(src, "EbsInfo: &ec2.EbsInfo{\n")
+		if info.EbsInfo.EbsOptimizedInfo != nil {
+			fmt.Fprintf(src, "EbsOptimizedInfo: &ec2.EbsOptimizedInfo{\n")
+			fmt.Fprintf(src, "BaselineBandwidthInMbps: aws.Int64(%d),\n", lo.FromPtr(info.EbsInfo.EbsOptimizedInfo.BaselineBandwidthInMbps))
+			fmt.Fprintf(src, "BaselineIops: aws.Int64(%d),\n", lo.FromPtr(info.EbsInfo.EbsOptimizedInfo.BaselineIops))
+			fmt.Fprintf(src, "BaselineThroughputInMBps: aws.Float64(%.2f),\n", lo.FromPtr(info.EbsInfo.EbsOptimizedInfo.BaselineThroughputInMBps))
+			fmt.Fprintf(src, "MaximumBandwidthInMbps: aws.Int64(%d),\n", lo.FromPtr(info.EbsInfo.EbsOptimizedInfo.MaximumBandwidthInMbps))
+			fmt.Fprintf(src, "MaximumIops: aws.Int64(%d),\n", lo.FromPtr(info.EbsInfo.EbsOptimizedInfo.MaximumIops))
+			fmt.Fprintf(src, "MaximumThroughputInMBps: aws.Float64(%.2f),\n", lo.FromPtr(info.EbsInfo.EbsOptimizedInfo.MaximumThroughputInMBps))
+			fmt.Fprintf(src, "},\n")
+		}
+		fmt.Fprintf(src, "EbsOptimizedSupport: aws.String(\"%s\"),\n", lo.FromPtr(info.EbsInfo.EbsOptimizedSupport))
+		fmt.Fprintf(src, "EncryptionSupport: aws.String(\"%s\"),\n", lo.FromPtr(info.EbsInfo.EncryptionSupport))
+		fmt.Fprintf(src, "NvmeSupport: aws.String(\"%s\"),\n", lo.FromPtr(info.EbsInfo.NvmeSupport))
+		fmt.Fprintf(src, "},\n")
+	}
 	if info.InferenceAcceleratorInfo != nil {
 		fmt.Fprintf(src, "InferenceAcceleratorInfo: &ec2.InferenceAcceleratorInfo{\n")
 		fmt.Fprintf(src, "Accelerators: []*ec2.InferenceDeviceInfo{\n")

diff --git a/hack/validation/labels.sh b/hack/validation/labels.sh
@@ -4,4 +4,4 @@
 
 # ## checking for restricted labels while filtering out well known labels
 yq eval '.spec.versions[0].schema.openAPIV3Schema.properties.spec.properties.template.properties.metadata.properties.labels.x-kubernetes-validations += [
-    {"message": "label domain \"karpenter.k8s.aws\" is restricted", "rule": "self.all(x, x in [\"karpenter.k8s.aws/instance-encryption-in-transit-supported\", \"karpenter.k8s.aws/instance-category\", \"karpenter.k8s.aws/instance-hypervisor\", \"karpenter.k8s.aws/instance-family\", \"karpenter.k8s.aws/instance-generation\", \"karpenter.k8s.aws/instance-local-nvme\", \"karpenter.k8s.aws/instance-size\", \"karpenter.k8s.aws/instance-cpu\",\"karpenter.k8s.aws/instance-cpu-manufacturer\",\"karpenter.k8s.aws/instance-memory\", \"karpenter.k8s.aws/instance-network-bandwidth\", \"karpenter.k8s.aws/instance-gpu-name\", \"karpenter.k8s.aws/instance-gpu-manufacturer\", \"karpenter.k8s.aws/instance-gpu-count\", \"karpenter.k8s.aws/instance-gpu-memory\", \"karpenter.k8s.aws/instance-accelerator-name\", \"karpenter.k8s.aws/instance-accelerator-manufacturer\", \"karpenter.k8s.aws/instance-accelerator-count\"] || !x.find(\"^([^/]+)\").endsWith(\"karpenter.k8s.aws\"))"}]' -i pkg/apis/crds/karpenter.sh_nodepools.yaml 
+    {"message": "label domain \"karpenter.k8s.aws\" is restricted", "rule": "self.all(x, x in [\"karpenter.k8s.aws/instance-encryption-in-transit-supported\", \"karpenter.k8s.aws/instance-category\", \"karpenter.k8s.aws/instance-hypervisor\", \"karpenter.k8s.aws/instance-family\", \"karpenter.k8s.aws/instance-generation\", \"karpenter.k8s.aws/instance-local-nvme\", \"karpenter.k8s.aws/instance-size\", \"karpenter.k8s.aws/instance-cpu\",\"karpenter.k8s.aws/instance-cpu-manufacturer\",\"karpenter.k8s.aws/instance-memory\", \"karpenter.k8s.aws/instance-ebs-bandwidth\", \"karpenter.k8s.aws/instance-network-bandwidth\", \"karpenter.k8s.aws/instance-gpu-name\", \"karpenter.k8s.aws/instance-gpu-manufacturer\", \"karpenter.k8s.aws/instance-gpu-count\", \"karpenter.k8s.aws/instance-gpu-memory\", \"karpenter.k8s.aws/instance-accelerator-name\", \"karpenter.k8s.aws/instance-accelerator-manufacturer\", \"karpenter.k8s.aws/instance-accelerator-count\"] || !x.find(\"^([^/]+)\").endsWith(\"karpenter.k8s.aws\"))"}]' -i pkg/apis/crds/karpenter.sh_nodepools.yaml 
diff --git a/hack/validation/requirements.sh b/hack/validation/requirements.sh
@@ -4,9 +4,9 @@
 
 ## checking for restricted labels while filtering out well known labels
 yq eval '.spec.versions[0].schema.openAPIV3Schema.properties.spec.properties.requirements.items.properties.key.x-kubernetes-validations += [
-    {"message": "label domain \"karpenter.k8s.aws\" is restricted", "rule": "self in [\"karpenter.k8s.aws/instance-encryption-in-transit-supported\", \"karpenter.k8s.aws/instance-category\", \"karpenter.k8s.aws/instance-hypervisor\", \"karpenter.k8s.aws/instance-family\", \"karpenter.k8s.aws/instance-generation\", \"karpenter.k8s.aws/instance-local-nvme\", \"karpenter.k8s.aws/instance-size\", \"karpenter.k8s.aws/instance-cpu\",\"karpenter.k8s.aws/instance-cpu-manufacturer\",\"karpenter.k8s.aws/instance-memory\", \"karpenter.k8s.aws/instance-network-bandwidth\", \"karpenter.k8s.aws/instance-gpu-name\", \"karpenter.k8s.aws/instance-gpu-manufacturer\", \"karpenter.k8s.aws/instance-gpu-count\", \"karpenter.k8s.aws/instance-gpu-memory\", \"karpenter.k8s.aws/instance-accelerator-name\", \"karpenter.k8s.aws/instance-accelerator-manufacturer\", \"karpenter.k8s.aws/instance-accelerator-count\"] || !self.find(\"^([^/]+)\").endsWith(\"karpenter.k8s.aws\")"}]' -i pkg/apis/crds/karpenter.sh_nodeclaims.yaml 
+    {"message": "label domain \"karpenter.k8s.aws\" is restricted", "rule": "self in [\"karpenter.k8s.aws/instance-encryption-in-transit-supported\", \"karpenter.k8s.aws/instance-category\", \"karpenter.k8s.aws/instance-hypervisor\", \"karpenter.k8s.aws/instance-family\", \"karpenter.k8s.aws/instance-generation\", \"karpenter.k8s.aws/instance-local-nvme\", \"karpenter.k8s.aws/instance-size\", \"karpenter.k8s.aws/instance-cpu\",\"karpenter.k8s.aws/instance-cpu-manufacturer\",\"karpenter.k8s.aws/instance-memory\", \"karpenter.k8s.aws/instance-ebs-bandwidth\", \"karpenter.k8s.aws/instance-network-bandwidth\", \"karpenter.k8s.aws/instance-gpu-name\", \"karpenter.k8s.aws/instance-gpu-manufacturer\", \"karpenter.k8s.aws/instance-gpu-count\", \"karpenter.k8s.aws/instance-gpu-memory\", \"karpenter.k8s.aws/instance-accelerator-name\", \"karpenter.k8s.aws/instance-accelerator-manufacturer\", \"karpenter.k8s.aws/instance-accelerator-count\"] || !self.find(\"^([^/]+)\").endsWith(\"karpenter.k8s.aws\")"}]' -i pkg/apis/crds/karpenter.sh_nodeclaims.yaml 
 # # Adding validation for nodepool
 
 # ## checking for restricted labels while filtering out well known labels
 yq eval '.spec.versions[0].schema.openAPIV3Schema.properties.spec.properties.template.properties.spec.properties.requirements.items.properties.key.x-kubernetes-validations  += [
-    {"message": "label domain \"karpenter.k8s.aws\" is restricted", "rule": "self in [\"karpenter.k8s.aws/instance-encryption-in-transit-supported\", \"karpenter.k8s.aws/instance-category\", \"karpenter.k8s.aws/instance-hypervisor\", \"karpenter.k8s.aws/instance-family\", \"karpenter.k8s.aws/instance-generation\", \"karpenter.k8s.aws/instance-local-nvme\", \"karpenter.k8s.aws/instance-size\", \"karpenter.k8s.aws/instance-cpu\",\"karpenter.k8s.aws/instance-cpu-manufacturer\",\"karpenter.k8s.aws/instance-memory\", \"karpenter.k8s.aws/instance-network-bandwidth\", \"karpenter.k8s.aws/instance-gpu-name\", \"karpenter.k8s.aws/instance-gpu-manufacturer\", \"karpenter.k8s.aws/instance-gpu-count\", \"karpenter.k8s.aws/instance-gpu-memory\", \"karpenter.k8s.aws/instance-accelerator-name\", \"karpenter.k8s.aws/instance-accelerator-manufacturer\", \"karpenter.k8s.aws/instance-accelerator-count\"] || !self.find(\"^([^/]+)\").endsWith(\"karpenter.k8s.aws\")"}]' -i pkg/apis/crds/karpenter.sh_nodepools.yaml 
+    {"message": "label domain \"karpenter.k8s.aws\" is restricted", "rule": "self in [\"karpenter.k8s.aws/instance-encryption-in-transit-supported\", \"karpenter.k8s.aws/instance-category\", \"karpenter.k8s.aws/instance-hypervisor\", \"karpenter.k8s.aws/instance-family\", \"karpenter.k8s.aws/instance-generation\", \"karpenter.k8s.aws/instance-local-nvme\", \"karpenter.k8s.aws/instance-size\", \"karpenter.k8s.aws/instance-cpu\",\"karpenter.k8s.aws/instance-cpu-manufacturer\",\"karpenter.k8s.aws/instance-memory\", \"karpenter.k8s.aws/instance-ebs-bandwidth\", \"karpenter.k8s.aws/instance-network-bandwidth\", \"karpenter.k8s.aws/instance-gpu-name\", \"karpenter.k8s.aws/instance-gpu-manufacturer\", \"karpenter.k8s.aws/instance-gpu-count\", \"karpenter.k8s.aws/instance-gpu-memory\", \"karpenter.k8s.aws/instance-accelerator-name\", \"karpenter.k8s.aws/instance-accelerator-manufacturer\", \"karpenter.k8s.aws/instance-accelerator-count\"] || !self.find(\"^([^/]+)\").endsWith(\"karpenter.k8s.aws\")"}]' -i pkg/apis/crds/karpenter.sh_nodepools.yaml 
diff --git a/pkg/apis/crds/karpenter.sh_nodeclaims.yaml b/pkg/apis/crds/karpenter.sh_nodeclaims.yaml
@@ -220,7 +220,7 @@ spec:
                           - message: label "kubernetes.io/hostname" is restricted
                             rule: self != "kubernetes.io/hostname"
                           - message: label domain "karpenter.k8s.aws" is restricted
-                            rule: self in ["karpenter.k8s.aws/instance-encryption-in-transit-supported", "karpenter.k8s.aws/instance-category", "karpenter.k8s.aws/instance-hypervisor", "karpenter.k8s.aws/instance-family", "karpenter.k8s.aws/instance-generation", "karpenter.k8s.aws/instance-local-nvme", "karpenter.k8s.aws/instance-size", "karpenter.k8s.aws/instance-cpu","karpenter.k8s.aws/instance-cpu-manufacturer","karpenter.k8s.aws/instance-memory", "karpenter.k8s.aws/instance-network-bandwidth", "karpenter.k8s.aws/instance-gpu-name", "karpenter.k8s.aws/instance-gpu-manufacturer", "karpenter.k8s.aws/instance-gpu-count", "karpenter.k8s.aws/instance-gpu-memory", "karpenter.k8s.aws/instance-accelerator-name", "karpenter.k8s.aws/instance-accelerator-manufacturer", "karpenter.k8s.aws/instance-accelerator-count"] || !self.find("^([^/]+)").endsWith("karpenter.k8s.aws")
+                            rule: self in ["karpenter.k8s.aws/instance-encryption-in-transit-supported", "karpenter.k8s.aws/instance-category", "karpenter.k8s.aws/instance-hypervisor", "karpenter.k8s.aws/instance-family", "karpenter.k8s.aws/instance-generation", "karpenter.k8s.aws/instance-local-nvme", "karpenter.k8s.aws/instance-size", "karpenter.k8s.aws/instance-cpu","karpenter.k8s.aws/instance-cpu-manufacturer","karpenter.k8s.aws/instance-memory", "karpenter.k8s.aws/instance-ebs-bandwidth", "karpenter.k8s.aws/instance-network-bandwidth", "karpenter.k8s.aws/instance-gpu-name", "karpenter.k8s.aws/instance-gpu-manufacturer", "karpenter.k8s.aws/instance-gpu-count", "karpenter.k8s.aws/instance-gpu-memory", "karpenter.k8s.aws/instance-accelerator-name", "karpenter.k8s.aws/instance-accelerator-manufacturer", "karpenter.k8s.aws/instance-accelerator-count"] || !self.find("^([^/]+)").endsWith("karpenter.k8s.aws")
                       minValues:
                         description: |-
                           This field is ALPHA and can be dropped or replaced at any time

diff --git a/pkg/apis/crds/karpenter.sh_nodepools.yaml b/pkg/apis/crds/karpenter.sh_nodepools.yaml
@@ -190,7 +190,7 @@ spec:
                             - message: label "kubernetes.io/hostname" is restricted
                               rule: self.all(x, x != "kubernetes.io/hostname")
                             - message: label domain "karpenter.k8s.aws" is restricted
-                              rule: self.all(x, x in ["karpenter.k8s.aws/instance-encryption-in-transit-supported", "karpenter.k8s.aws/instance-category", "karpenter.k8s.aws/instance-hypervisor", "karpenter.k8s.aws/instance-family", "karpenter.k8s.aws/instance-generation", "karpenter.k8s.aws/instance-local-nvme", "karpenter.k8s.aws/instance-size", "karpenter.k8s.aws/instance-cpu","karpenter.k8s.aws/instance-cpu-manufacturer","karpenter.k8s.aws/instance-memory", "karpenter.k8s.aws/instance-network-bandwidth", "karpenter.k8s.aws/instance-gpu-name", "karpenter.k8s.aws/instance-gpu-manufacturer", "karpenter.k8s.aws/instance-gpu-count", "karpenter.k8s.aws/instance-gpu-memory", "karpenter.k8s.aws/instance-accelerator-name", "karpenter.k8s.aws/instance-accelerator-manufacturer", "karpenter.k8s.aws/instance-accelerator-count"] || !x.find("^([^/]+)").endsWith("karpenter.k8s.aws"))
+                              rule: self.all(x, x in ["karpenter.k8s.aws/instance-encryption-in-transit-supported", "karpenter.k8s.aws/instance-category", "karpenter.k8s.aws/instance-hypervisor", "karpenter.k8s.aws/instance-family", "karpenter.k8s.aws/instance-generation", "karpenter.k8s.aws/instance-local-nvme", "karpenter.k8s.aws/instance-size", "karpenter.k8s.aws/instance-cpu","karpenter.k8s.aws/instance-cpu-manufacturer","karpenter.k8s.aws/instance-memory", "karpenter.k8s.aws/instance-ebs-bandwidth", "karpenter.k8s.aws/instance-network-bandwidth", "karpenter.k8s.aws/instance-gpu-name", "karpenter.k8s.aws/instance-gpu-manufacturer", "karpenter.k8s.aws/instance-gpu-count", "karpenter.k8s.aws/instance-gpu-memory", "karpenter.k8s.aws/instance-accelerator-name", "karpenter.k8s.aws/instance-accelerator-manufacturer", "karpenter.k8s.aws/instance-accelerator-count"] || !x.find("^([^/]+)").endsWith("karpenter.k8s.aws"))
                       type: object
                     spec:
                       description: NodeClaimSpec describes the desired state of the NodeClaim
@@ -348,7 +348,7 @@ spec:
                                   - message: label "kubernetes.io/hostname" is restricted
                                     rule: self != "kubernetes.io/hostname"
                                   - message: label domain "karpenter.k8s.aws" is restricted
-                                    rule: self in ["karpenter.k8s.aws/instance-encryption-in-transit-supported", "karpenter.k8s.aws/instance-category", "karpenter.k8s.aws/instance-hypervisor", "karpenter.k8s.aws/instance-family", "karpenter.k8s.aws/instance-generation", "karpenter.k8s.aws/instance-local-nvme", "karpenter.k8s.aws/instance-size", "karpenter.k8s.aws/instance-cpu","karpenter.k8s.aws/instance-cpu-manufacturer","karpenter.k8s.aws/instance-memory", "karpenter.k8s.aws/instance-network-bandwidth", "karpenter.k8s.aws/instance-gpu-name", "karpenter.k8s.aws/instance-gpu-manufacturer", "karpenter.k8s.aws/instance-gpu-count", "karpenter.k8s.aws/instance-gpu-memory", "karpenter.k8s.aws/instance-accelerator-name", "karpenter.k8s.aws/instance-accelerator-manufacturer", "karpenter.k8s.aws/instance-accelerator-count"] || !self.find("^([^/]+)").endsWith("karpenter.k8s.aws")
+                                    rule: self in ["karpenter.k8s.aws/instance-encryption-in-transit-supported", "karpenter.k8s.aws/instance-category", "karpenter.k8s.aws/instance-hypervisor", "karpenter.k8s.aws/instance-family", "karpenter.k8s.aws/instance-generation", "karpenter.k8s.aws/instance-local-nvme", "karpenter.k8s.aws/instance-size", "karpenter.k8s.aws/instance-cpu","karpenter.k8s.aws/instance-cpu-manufacturer","karpenter.k8s.aws/instance-memory", "karpenter.k8s.aws/instance-ebs-bandwidth", "karpenter.k8s.aws/instance-network-bandwidth", "karpenter.k8s.aws/instance-gpu-name", "karpenter.k8s.aws/instance-gpu-manufacturer", "karpenter.k8s.aws/instance-gpu-count", "karpenter.k8s.aws/instance-gpu-memory", "karpenter.k8s.aws/instance-accelerator-name", "karpenter.k8s.aws/instance-accelerator-manufacturer", "karpenter.k8s.aws/instance-accelerator-count"] || !self.find("^([^/]+)").endsWith("karpenter.k8s.aws")
                               minValues:
                                 description: |-
                                   This field is ALPHA and can be dropped or replaced at any time

diff --git a/pkg/apis/v1beta1/labels.go b/pkg/apis/v1beta1/labels.go
@@ -37,6 +37,7 @@ func init() {
 		LabelInstanceCPU,
 		LabelInstanceCPUManufacturer,
 		LabelInstanceMemory,
+		LabelInstanceEBSBandwidth,
 		LabelInstanceNetworkBandwidth,
 		LabelInstanceGPUName,
 		LabelInstanceGPUManufacturer,
@@ -103,6 +104,7 @@ var (
 	LabelInstanceCPU                          = Group + "/instance-cpu"
 	LabelInstanceCPUManufacturer              = Group + "/instance-cpu-manufacturer"
 	LabelInstanceMemory                       = Group + "/instance-memory"
+	LabelInstanceEBSBandwidth                 = Group + "/instance-ebs-bandwidth"
 	LabelInstanceNetworkBandwidth             = Group + "/instance-network-bandwidth"
 	LabelInstanceGPUName                      = Group + "/instance-gpu-name"
 	LabelInstanceGPUManufacturer              = Group + "/instance-gpu-manufacturer"