kubernetes/kops · error

error enabling metrics collection for AutoscalingGroup: %v

Error message

error enabling metrics collection for AutoscalingGroup: %v

What it means

RenderAWS failed to call EnableMetricsCollection on the newly created AutoScalingGroup: the ASG itself was created, but detailed monitoring could not be enabled — typically an IAM permission (autoscaling:EnableMetricsCollection), throttling, or an invalid metrics/granularity combination.

Source

Thrown at upup/pkg/fi/cloudup/awstasks/autoscalinggroup.go:440

		// @step: attempt to create the autoscaling group for us
		if _, err := t.Cloud.Autoscaling().CreateAutoScalingGroup(ctx, request); err != nil {
			code := awsup.AWSErrorCode(err)
			message := awsup.AWSErrorMessage(err)
			if code == "ValidationError" && strings.Contains(message, "Invalid IAM Instance Profile name") {
				klog.V(4).Infof("error creating AutoscalingGroup: %s", message)
				return fi.NewTryAgainLaterError("waiting for the IAM Instance Profile to be propagated")
			}
			return fmt.Errorf("error creating AutoScalingGroup: %s", message)
		}

		// @step: attempt to enable the metrics for us
		if _, err := t.Cloud.Autoscaling().EnableMetricsCollection(ctx, &autoscaling.EnableMetricsCollectionInput{
			AutoScalingGroupName: e.Name,
			Granularity:          e.Granularity,
			Metrics:              e.Metrics,
		}); err != nil {
			return fmt.Errorf("error enabling metrics collection for AutoscalingGroup: %v", err)
		}

		if len(*e.SuspendProcesses) > 0 {
			processQuery := &autoscaling.SuspendProcessesInput{}
			processQuery.AutoScalingGroupName = e.Name
			processQuery.ScalingProcesses = *e.SuspendProcesses

			if _, err := t.Cloud.Autoscaling().SuspendProcesses(ctx, processQuery); err != nil {
				return fmt.Errorf("error suspending processes: %v", err)
			}
		}

	} else {
		// @logic: else we have found a autoscaling group and we need to evaluate the difference
		request := &autoscaling.UpdateAutoScalingGroupInput{
			AutoScalingGroupName: e.Name,
		}

View on GitHub (pinned to 4c8573c808)

Solutions

1. Check `metrics:` and granularity in the instance group spec; remove unsupported metric names (valid: e.g. GroupMinSize, GroupMaxSize, GroupDesiredCapacity...)
2. Ensure granularity is '1Minute'
3. Re-run `kops update cluster` — the ASG exists, only metrics enablement failed
4. Verify IAM permissions for autoscaling:EnableMetricsCollection

Example fix

// before
metrics: [GroupMinSize, GroupTotalInstances, BogusMetric]
// after
metrics: [GroupMinSize, GroupMaxSize, GroupDesiredCapacity, GroupInServiceInstances]

valid := map[string]bool{"GroupMinSize":true,"GroupMaxSize":true,"GroupDesiredCapacity":true,"GroupInServiceInstances":true,"GroupPendingInstances":true,"GroupStandbyInstances":true,"GroupTerminatingInstances":true,"GroupTotalInstances":true}
for _, m := range ig.Spec.Metrics {
  if !valid[m] { return fmt.Errorf("unsupported ASG metric %q", m) }
}
if ig.Spec.Granularity != "" && ig.Spec.Granularity != "1Minute" { return errors.New("granularity must be 1Minute") }

When it happens

Trigger: CreateAutoScalingGroup succeeded, but the follow-up EnableMetricsCollection input is rejected — e.g. invalid Granularity (must be '1Minute'), invalid metric names, or transient AWS errors.

Common situations: Custom metrics list containing a misspelled/unsupported metric, or empty/incorrect granularity after editing instance group spec `metrics:` fields.

Related errors

• DIGITALOCEAN_ACCESS_TOKEN is required
• the image for the hook exec action not set
• IP version is incorrect
• ErrAlreadyExists
• DIGITALOCEAN_ACCESS_TOKEN is required

AI-assisted analysis of kubernetes/kops@4c8573c808 (2026-09-05). Data as JSON: /api/errors/0238a6e96f81847b. Report an issue: GitHub.