kubernetes/kops · error

unable to find VM Scale Sets: %s

Error message

unable to find VM Scale Sets: %s

What it means

GetCloudGroups fails when c.vmscaleSetsClient.List cannot enumerate VM Scale Sets in the cluster resource group. Cloud groups (instance groups mapped to VMSS) are required for rolling updates and drain logic, so the operation aborts.

Source

Thrown at upup/pkg/fi/cloudup/azure/status.go:131

}

// GetCloudGroups returns Cloud Instance Groups for the cluster
// by querying Azure.
func (c *azureCloudImplementation) GetCloudGroups(
	cluster *kops.Cluster,
	instancegroups []*kops.InstanceGroup,
	warnUnmatched bool,
	nodes []v1.Node,
) (map[string]*cloudinstances.CloudInstanceGroup, error) {
	igsByName, err := keyedByName(instancegroups, cluster.Name)
	if err != nil {
		return nil, err
	}

	ctx := context.TODO()
	vmsses, err := c.vmscaleSetsClient.List(ctx, cluster.AzureResourceGroupName())
	if err != nil {
		return nil, fmt.Errorf("unable to find VM Scale Sets: %s", err)
	}

	nodeMap := cloudinstances.GetNodeMap(nodes, cluster)

	groups := make(map[string]*cloudinstances.CloudInstanceGroup)
	for _, vmss := range vmsses {
		if !isOwnedByCluster(vmss, cluster.Name) {
			continue
		}

		ig, ok := igsByName[*vmss.Name]
		if !ok {
			if warnUnmatched {
				klog.Warningf("Found VM Scale Set with no corresponding instance group %q", *vmss.Name)
			}
			continue
		}

View on GitHub (pinned to 4c8573c808)

Solutions

1. Grant the identity Microsoft.Compute/virtualMachineScaleSets/read (Reader/Contributor) on the resource group
2. Confirm the cluster's AzureResourceGroupName is correct and exists
3. Retry on throttling (429) honoring Retry-After
4. Check Azure status for ARM outages if errors are transient 5xx

if _, err := rgClient.Get(ctx, cluster.AzureResourceGroupName(), nil); err != nil {
	return fmt.Errorf("cluster resource group %q not found: %w", cluster.AzureResourceGroupName(), err)
}

var respErr *azcore.ResponseError
if errors.As(err, &respErr) {
	if respErr.StatusCode == http.StatusForbidden {
		// missing Microsoft.Compute/virtualMachineScaleSets/read
	}
}

groups, err := cloud.GetCloudGroups(cluster, groups, nodes, instanceGroups)
if err != nil {
	var respErr *azcore.ResponseError
	if errors.As(err, &respErr) && respErr.StatusCode == 429 {
		return retryAfterBackoff(ctx, respErr.RetryAfter)
	}
	return err
}

When it happens

Trigger: vmscaleSetsClient.List(ctx, cluster.AzureResourceGroupName()) returns error: missing Microsoft.Compute/virtualMachineScaleSets/read permission, resource group not found beyond the client's tolerance, throttling, or transient ARM errors.

Common situations: kops rolling-update cluster on Azure with an identity lacking Compute read permissions; resource group renamed/deleted; subscription mismatch; Azure regional outage.

Related errors

• expected exactly one subnet for InstanceGroup %q; subnets wa
• unexpected subnet type: for InstanceGroup %q; type was %s
• instance group must have the same min and max size in Azure,
• malformed format of image urn: %s
• creating VMSS VMs client: %w

AI-assisted analysis of kubernetes/kops@4c8573c808 (2026-09-05). Data as JSON: /api/errors/69d2c30448f1deeb. Report an issue: GitHub.