Remove AWS options from CLI args

pkg/operator/operator.go

@@ -96,13 +96,14 @@ func NewOperator() (context.Context, *Operator) {
 
 	// Settings
 	settingsStore := settingsstore.WatchSettingsOrDie(ctx, kubernetesInterface, configMapWatcher, apis.Settings.List()...)
-	ctx = settingsStore.InjectSettings(ctx)
 
 	// Logging
 	logger := NewLogger(ctx, component, config, configMapWatcher)
 	ctx = logging.WithLogger(ctx, logger)
 
+	// Inject settings after starting the ConfigMapWatcher
 	lo.Must0(configMapWatcher.Start(ctx.Done()))
+	ctx = settingsStore.InjectSettings(ctx)
 
 	// Manager
 	manager, err := controllerruntime.NewManager(config, controllerruntime.Options{

pkg/operator/options/options.go

@@ -17,13 +17,9 @@ package options
 import (
 	"errors"
 	"flag"
-	"fmt"
-	"net/url"
 	"os"
 	"runtime/debug"
 
-	"go.uber.org/multierr"
-
 	"github.com/aws/karpenter-core/pkg/utils/env"
 )
 
@@ -47,15 +43,6 @@ type Options struct {
 	EnableProfiling      bool
 	EnableLeaderElection bool
 	MemoryLimit          int64
-	// AWS Specific
-	ClusterName               string
-	ClusterEndpoint           string
-	VMMemoryOverhead          float64
-	AWSNodeNameConvention     string
-	AWSENILimitedPodDensity   bool
-	AWSDefaultInstanceProfile string
-	AWSEnablePodENI           bool
-	AWSIsolatedVPC            bool
 }
 
 // New creates an Options struct and registers CLI flags and environment variables to fill-in the Options struct fields
@@ -74,15 +61,6 @@ func New() *Options {
 	f.BoolVar(&opts.EnableProfiling, "enable-profiling", env.WithDefaultBool("ENABLE_PROFILING", false), "Enable the profiling on the metric endpoint")
 	f.BoolVar(&opts.EnableLeaderElection, "leader-elect", env.WithDefaultBool("LEADER_ELECT", true), "Start leader election client and gain leadership before executing the main loop. Enable this when running replicated components for high availability.")
 	f.Int64Var(&opts.MemoryLimit, "memory-limit", env.WithDefaultInt64("MEMORY_LIMIT", -1), "Memory limit on the container running the controller. The GC soft memory limit is set to 90% of this value.")
-	// AWS Specific
-	f.StringVar(&opts.ClusterName, "cluster-name", env.WithDefaultString("CLUSTER_NAME", ""), "The kubernetes cluster name for resource discovery")
-	f.StringVar(&opts.ClusterEndpoint, "cluster-endpoint", env.WithDefaultString("CLUSTER_ENDPOINT", ""), "The external kubernetes cluster endpoint for new nodes to connect with")
-	f.Float64Var(&opts.VMMemoryOverhead, "vm-memory-overhead", env.WithDefaultFloat64("VM_MEMORY_OVERHEAD", 0.075), "The VM memory overhead as a percent that will be subtracted from the total memory for all instance types")
-	f.StringVar(&opts.AWSNodeNameConvention, "aws-node-name-convention", env.WithDefaultString("AWS_NODE_NAME_CONVENTION", string(IPName)), "The node naming convention used by the AWS cloud provider. DEPRECATION WARNING: this field may be deprecated at any time")
-	f.BoolVar(&opts.AWSENILimitedPodDensity, "aws-eni-limited-pod-density", env.WithDefaultBool("AWS_ENI_LIMITED_POD_DENSITY", true), "Indicates whether new nodes should use ENI-based pod density. DEPRECATED: Use `.spec.kubeletConfiguration.maxPods` to set pod density on a per-provisioner basis")
-	f.StringVar(&opts.AWSDefaultInstanceProfile, "aws-default-instance-profile", env.WithDefaultString("AWS_DEFAULT_INSTANCE_PROFILE", ""), "The default instance profile to use when provisioning nodes in AWS")
-	f.BoolVar(&opts.AWSEnablePodENI, "aws-enable-pod-eni", env.WithDefaultBool("AWS_ENABLE_POD_ENI", false), "If true then instances that support pod ENI will report a vpc.amazonaws.com/pod-eni resource")
-	f.BoolVar(&opts.AWSIsolatedVPC, "aws-isolated-vpc", env.WithDefaultBool("AWS_ISOLATED_VPC", false), "If true then assume we can't reach AWS services which don't have a VPC endpoint. This also has the effect of disabling look-ups to the AWS pricing endpoint.")
 
 	if opts.MemoryLimit > 0 {
 		newLimit := int64(float64(opts.MemoryLimit) * 0.9)
@@ -102,34 +80,5 @@ func (o *Options) MustParse() *Options {
 	if err != nil {
 		panic(err)
 	}
-	if err := o.Validate(); err != nil {
-		panic(err)
-	}
 	return o
 }
-
-func (o Options) Validate() (err error) {
-	err = multierr.Append(err, o.validateEndpoint())
-	if o.ClusterName == "" {
-		err = multierr.Append(err, fmt.Errorf("CLUSTER_NAME is required"))
-	}
-	awsNodeNameConvention := AWSNodeNameConvention(o.AWSNodeNameConvention)
-	if awsNodeNameConvention != IPName && awsNodeNameConvention != ResourceName {
-		err = multierr.Append(err, fmt.Errorf("aws-node-name-convention may only be either ip-name or resource-name"))
-	}
-	return err
-}
-
-func (o Options) validateEndpoint() error {
-	endpoint, err := url.Parse(o.ClusterEndpoint)
-	// url.Parse() will accept a lot of input without error; make
-	// sure it's a real URL
-	if err != nil || !endpoint.IsAbs() || endpoint.Hostname() == "" {
-		return fmt.Errorf("\"%s\" not a valid CLUSTER_ENDPOINT URL", o.ClusterEndpoint)
-	}
-	return nil
-}
-
-func (o Options) GetAWSNodeNameConvention() AWSNodeNameConvention {
-	return AWSNodeNameConvention(o.AWSNodeNameConvention)
-}