kubernetes/kops · error

error initializing instance selector: %v

Error message

error initializing instance selector: %v

What it means

This wraps a failure from selector.New(ctx, cloud.Config()), which initializes the kOps instance-type selector by fetching EC2 instance type/price data (via the AWS pricing/EC2 APIs). If that discovery fails, the selector cannot be constructed and the command aborts.

Source

Thrown at cmd/kops/toolbox_instance-selector.go:276

			return err
		}
		igSubnets = userSubnets
	}

	zones := []string{}
	for _, igSubnet := range igSubnets {
		zones = append(zones, strings.ReplaceAll(igSubnet, "utility-", ""))
	}

	tags := map[string]string{"KubernetesCluster": options.ClusterName}
	cloud, err := awsup.NewAWSCloud(region, tags)
	if err != nil {
		return fmt.Errorf("error initializing AWS client: %v", err)
	}

	instanceSelector, err := selector.New(ctx, cloud.Config())
	if err != nil {
		return fmt.Errorf("error initializing instance selector: %v", err)
	}
	igCount := options.InstanceGroupCount
	filters := getFilters(commandline, region, zones)
	mutatedFilters := filters
	if commandline.Flags[instanceGroupCount] != nil || filters.Flexible != nil {
		if filters.VCpusToMemoryRatio == nil {
			defaultStartRatio := float64(2.0)
			mutatedFilters.VCpusToMemoryRatio = &defaultStartRatio
		}
	}

	newInstanceGroups := []*kops.InstanceGroup{}

	for i := 0; i < igCount; i++ {
		igNameForRun := options.InstanceGroupName
		if igCount != 1 {
			igNameForRun = fmt.Sprintf("%s%d", options.InstanceGroupName, i+1)
		}

View on GitHub (pinned to 4c8573c808)

Solutions

1. Grant the caller IAM permissions ec2:DescribeInstanceTypes and access to the AWS Pricing API.
2. Retry after a short delay if throttling (RequestLimitExceeded) is the underlying cause.
3. Confirm network access to ec2 and pricing endpoints for the region.
4. Upgrade kOps if the region is newly launched and not yet known to this build.

Example fix

// IAM policy: allow selector initialization
// before
{"Effect": "Deny", "Action": ["ec2:Describe*"], "Resource": "*"}
// after
{"Effect": "Allow", "Action": ["ec2:DescribeInstanceTypes", "ec2:DescribeAvailabilityZones"], "Resource": "*"}

// Pre-flight IAM check
// aws ec2 describe-instance-types --region $REGION --max-items 1
// exit code != 0 => permissions/network problem

var selectorInstance *selector.InstanceSelector
for attempt := 0; attempt < 3; attempt++ {
	selectorInstance, err = selector.New(ctx, cloud.Config())
	if err == nil { break }
	time.Sleep(time.Duration(1<<attempt) * time.Second)
}

When it happens

Trigger: selector.New fails while enumerating EC2 instance types or pricing for the region: API throttling, denied permissions on ec2:DescribeInstanceTypes or pricing API, network failure, or unsupported/empty region data.

Common situations: IAM policy lacking ec2:DescribeInstanceTypes; running in a region where pricing data is unavailable; transient AWS API throttling on large accounts; network egress blocked to pricing endpoints.

Related errors

• cannot select instance types from non-aws cluster
• error finding matching instance types: %w
• error querying EC2 for user metadata for instance %q: %v
• error deleting tags on %v: %v
• DIGITALOCEAN_ACCESS_TOKEN is required

AI-assisted analysis of kubernetes/kops@4c8573c808 (2026-09-05). Data as JSON: /api/errors/af681680e096eb1b. Report an issue: GitHub.