operators: infrastructure autodiscovery (#1958)

* helm: configure GCP cloud controller manager to search in all zones of a region

See also: d716fdd452/providers/gce/gce.go (L376-L380)

* operators: add nodeGroupName to ScalingGroup CRD

NodeGroupName is the human friendly name of the node group that will be exposed to customers via the Constellation config in the future.

* operators: support simple executor / scheduler to reconcile on non-k8s resources

* operators: add new return type for ListScalingGroups to support arbitrary node groups

* operators: ListScalingGroups should return additionally created node groups on AWS

* operators: ListScalingGroups should return additionally created node groups on Azure

* operators: ListScalingGroups should return additionally created node groups on GCP

* operators: ListScalingGroups should return additionally created node groups on unsupported CSPs

* operators: implement external scaling group reconciler

This controller scans the cloud provider infrastructure and changes k8s resources accordingly.
It creates ScaleSet resources when new node groups are created and deletes them if the node groups are removed.

* operators: no longer create scale sets when the operator starts

In the future, scale sets are created dynamically.

* operators: watch for node join/leave events using a controller

* operators: deploy new controllers

* docs: update auto scaling documentation with support for node groups
This commit is contained in:
Malte Poll 2023-07-05 07:27:34 +02:00 committed by Adrian Stobbe
parent 10a540c290
commit 388ff011a3
36 changed files with 1836 additions and 232 deletions

View file

@ -19,6 +19,7 @@ go_library(
visibility = ["//operators/constellation-node-operator:__subpackages__"],
deps = [
"//operators/constellation-node-operator/api/v1alpha1",
"//operators/constellation-node-operator/internal/cloud/api",
"//operators/constellation-node-operator/internal/poller",
"@com_github_azure_azure_sdk_for_go_sdk_azcore//:azcore",
"@com_github_azure_azure_sdk_for_go_sdk_azcore//runtime",
@ -44,6 +45,7 @@ go_test(
embed = [":client"],
deps = [
"//operators/constellation-node-operator/api/v1alpha1",
"//operators/constellation-node-operator/internal/cloud/api",
"//operators/constellation-node-operator/internal/poller",
"@com_github_azure_azure_sdk_for_go_sdk_azcore//:azcore",
"@com_github_azure_azure_sdk_for_go_sdk_azcore//runtime",

View file

@ -13,6 +13,8 @@ import (
"github.com/Azure/azure-sdk-for-go/sdk/azcore/to"
"github.com/Azure/azure-sdk-for-go/sdk/resourcemanager/compute/armcompute/v4"
updatev1alpha1 "github.com/edgelesssys/constellation/v2/operators/constellation-node-operator/v2/api/v1alpha1"
cspapi "github.com/edgelesssys/constellation/v2/operators/constellation-node-operator/v2/internal/cloud/api"
)
// GetScalingGroupImage returns the image URI of the scaling group.
@ -79,34 +81,63 @@ func (c *Client) GetAutoscalingGroupName(scalingGroupID string) (string, error)
}
// ListScalingGroups retrieves a list of scaling groups for the cluster.
func (c *Client) ListScalingGroups(ctx context.Context, uid string) (controlPlaneGroupIDs []string, workerGroupIDs []string, err error) {
func (c *Client) ListScalingGroups(ctx context.Context, uid string) ([]cspapi.ScalingGroup, error) {
results := []cspapi.ScalingGroup{}
pager := c.scaleSetsAPI.NewListPager(c.config.ResourceGroup, nil)
for pager.More() {
page, err := pager.NextPage(ctx)
if err != nil {
return nil, nil, fmt.Errorf("paging scale sets: %w", err)
return nil, fmt.Errorf("paging scale sets: %w", err)
}
for _, scaleSet := range page.Value {
if scaleSet == nil || scaleSet.ID == nil {
continue
}
if scaleSet.Tags == nil || scaleSet.Tags["constellation-uid"] == nil || *scaleSet.Tags["constellation-uid"] != uid {
if scaleSet.Tags == nil ||
scaleSet.Tags["constellation-uid"] == nil ||
*scaleSet.Tags["constellation-uid"] != uid ||
scaleSet.Tags["constellation-role"] == nil {
continue
}
role := updatev1alpha1.NodeRoleFromString(*scaleSet.Tags["constellation-role"])
name, err := c.GetScalingGroupName(*scaleSet.ID)
if err != nil {
return nil, nil, fmt.Errorf("getting scaling group name: %w", err)
return nil, fmt.Errorf("getting scaling group name: %w", err)
}
switch *scaleSet.Tags["constellation-role"] {
case "control-plane", "controlplane":
controlPlaneGroupIDs = append(controlPlaneGroupIDs, *scaleSet.ID)
case "worker":
workerGroupIDs = append(workerGroupIDs, *scaleSet.ID)
var nodeGroupName string
if scaleSet.Tags["constellation-node-group"] != nil {
nodeGroupName = *scaleSet.Tags["constellation-node-group"]
}
// fallback for legacy clusters
// TODO(malt3): remove this fallback once we can assume all clusters have the correct labels
if nodeGroupName == "" {
switch role {
case updatev1alpha1.ControlPlaneRole:
nodeGroupName = "control_plane_default"
case updatev1alpha1.WorkerRole:
nodeGroupName = "worker_default"
}
}
autoscalerGroupName, err := c.GetAutoscalingGroupName(*scaleSet.ID)
if err != nil {
return nil, fmt.Errorf("getting autoscaling group name: %w", err)
}
results = append(results, cspapi.ScalingGroup{
Name: name,
NodeGroupName: nodeGroupName,
GroupID: *scaleSet.ID,
AutoscalingGroupName: autoscalerGroupName,
Role: role,
})
}
}
return controlPlaneGroupIDs, workerGroupIDs, nil
return results, nil
}
func imageReferenceFromImage(img string) *armcompute.ImageReference {

View file

@ -13,6 +13,7 @@ import (
"github.com/Azure/azure-sdk-for-go/sdk/azcore/to"
"github.com/Azure/azure-sdk-for-go/sdk/resourcemanager/compute/armcompute/v4"
cspapi "github.com/edgelesssys/constellation/v2/operators/constellation-node-operator/v2/internal/cloud/api"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
@ -184,41 +185,67 @@ func TestGetScalingGroupName(t *testing.T) {
func TestListScalingGroups(t *testing.T) {
testCases := map[string]struct {
scaleSet armcompute.VirtualMachineScaleSet
fetchPageErr error
wantControlPlanes []string
wantWorkers []string
wantErr bool
scaleSet armcompute.VirtualMachineScaleSet
fetchPageErr error
wantGroups []cspapi.ScalingGroup
wantErr bool
}{
"listing control-plane works": {
scaleSet: armcompute.VirtualMachineScaleSet{
ID: to.Ptr("/subscriptions/subscription-id/resourceGroups/resource-group/providers/Microsoft.Compute/virtualMachineScaleSets/constellation-scale-set-control-planes-uid"),
Name: to.Ptr("constellation-scale-set-control-planes-uid"),
ID: to.Ptr("/subscriptions/subscription-id/resourceGroups/resource-group/providers/Microsoft.Compute/virtualMachineScaleSets/constellation-scale-set-control-planes-uid"),
Tags: map[string]*string{
"constellation-uid": to.Ptr("uid"),
"constellation-role": to.Ptr("control-plane"),
},
},
wantControlPlanes: []string{"/subscriptions/subscription-id/resourceGroups/resource-group/providers/Microsoft.Compute/virtualMachineScaleSets/constellation-scale-set-control-planes-uid"},
wantGroups: []cspapi.ScalingGroup{
{
Name: "constellation-scale-set-control-planes-uid",
NodeGroupName: "control_plane_default",
GroupID: "/subscriptions/subscription-id/resourceGroups/resource-group/providers/Microsoft.Compute/virtualMachineScaleSets/constellation-scale-set-control-planes-uid",
AutoscalingGroupName: "constellation-scale-set-control-planes-uid",
Role: "ControlPlane",
},
},
},
"listing worker works": {
scaleSet: armcompute.VirtualMachineScaleSet{
ID: to.Ptr("/subscriptions/subscription-id/resourceGroups/resource-group/providers/Microsoft.Compute/virtualMachineScaleSets/constellation-scale-set-workers-uid"),
Name: to.Ptr("constellation-scale-set-workers-uid"),
ID: to.Ptr("/subscriptions/subscription-id/resourceGroups/resource-group/providers/Microsoft.Compute/virtualMachineScaleSets/constellation-scale-set-workers-uid"),
Tags: map[string]*string{
"constellation-uid": to.Ptr("uid"),
"constellation-role": to.Ptr("worker"),
},
},
wantWorkers: []string{"/subscriptions/subscription-id/resourceGroups/resource-group/providers/Microsoft.Compute/virtualMachineScaleSets/constellation-scale-set-workers-uid"},
wantGroups: []cspapi.ScalingGroup{
{
Name: "constellation-scale-set-workers-uid",
NodeGroupName: "worker_default",
GroupID: "/subscriptions/subscription-id/resourceGroups/resource-group/providers/Microsoft.Compute/virtualMachineScaleSets/constellation-scale-set-workers-uid",
AutoscalingGroupName: "constellation-scale-set-workers-uid",
Role: "Worker",
},
},
},
"listing is not dependent on resource name": {
scaleSet: armcompute.VirtualMachineScaleSet{
ID: to.Ptr("/subscriptions/subscription-id/resourceGroups/resource-group/providers/Microsoft.Compute/virtualMachineScaleSets/some-scale-set"),
Name: to.Ptr("foo-bar"),
ID: to.Ptr("/subscriptions/subscription-id/resourceGroups/resource-group/providers/Microsoft.Compute/virtualMachineScaleSets/some-scale-set"),
Tags: map[string]*string{
"constellation-uid": to.Ptr("uid"),
"constellation-role": to.Ptr("control-plane"),
},
},
wantControlPlanes: []string{"/subscriptions/subscription-id/resourceGroups/resource-group/providers/Microsoft.Compute/virtualMachineScaleSets/some-scale-set"},
wantGroups: []cspapi.ScalingGroup{
{
Name: "some-scale-set",
NodeGroupName: "control_plane_default",
GroupID: "/subscriptions/subscription-id/resourceGroups/resource-group/providers/Microsoft.Compute/virtualMachineScaleSets/some-scale-set",
AutoscalingGroupName: "some-scale-set",
Role: "ControlPlane",
},
},
},
"listing other works": {
scaleSet: armcompute.VirtualMachineScaleSet{
@ -245,14 +272,13 @@ func TestListScalingGroups(t *testing.T) {
},
},
}
gotControlPlanes, gotWorkers, err := client.ListScalingGroups(context.Background(), "uid")
gotGroups, err := client.ListScalingGroups(context.Background(), "uid")
if tc.wantErr {
assert.Error(err)
return
}
require.NoError(err)
assert.ElementsMatch(tc.wantControlPlanes, gotControlPlanes)
assert.ElementsMatch(tc.wantWorkers, gotWorkers)
assert.ElementsMatch(tc.wantGroups, gotGroups)
})
}
}