{AKS} Add support for control plane scaling profile (#9670)

Author: nethbtju

src/aks-preview/HISTORY.rst

@@ -11,6 +11,7 @@ To release a new version, please select a new version number (usually plus 1 to
 
 Pending
 +++++++
+* `az aks create`: Add `--control-plane-scaling-size` parameter to configure control plane scaling profile with available sizes 'H2', 'H4', and 'H8'.
 
 20.0.0b4
 +++++++
@@ -48,7 +49,7 @@ Pending
 19.0.0b28
 +++++++
 * Fix `match_condition` kwarg leaking to HTTP transport by overriding `put_mc` and `add_agentpool` to pass `if_match` / `if_none_match` directly to the vendored SDK. This change fixes the compatibility issue as azure-cli/acs module adopts TypeSpec emitted SDKs while azure-cli-extensions/aks-preview still uses the autorest emitted SDK.
-+ `az aks list-vm-skus`: New command to list available VM SKUs for AKS clusters in a given region.
+* `az aks list-vm-skus`: New command to list available VM SKUs for AKS clusters in a given region.
 * `az aks create/update`: Add `--enable-service-account-image-pull`, `--disable-service-account-image-pull`, and `--service-account-image-pull-default-managed-identity-id` parameters to manage service account based image pull settings.
 * `az aks list-vm-skus`: New command to list available VM SKUs for AKS clusters in a given region.
 * Add managed GPU enablement option to node pool property in `az aks nodepool add` and `az aks nodepool update`.

src/aks-preview/azext_aks_preview/_help.py

@@ -715,6 +715,13 @@
         - name: --enable-hosted-system
           type: bool
           short-summary: Create a cluster with fully hosted system components. This applies only when creating a new automatic cluster.
+        - name: --control-plane-scaling-size --cp-scaling-size
+          type: string
+          short-summary: (PREVIEW) The control plane scaling size for the cluster.
+          long-summary: |
+              Provides scaled and performance-guaranteed control plane capacity for AKS clusters.
+              Enables customers to select a control plane scaling size that delivers higher API server throughput,
+              increased etcd capacity, and faster pod scheduling rates. Available values are 'H2', 'H4', and 'H8'.
     examples:
         - name: Create a Kubernetes cluster with an existing SSH public key.
           text: az aks create -g MyResourceGroup -n MyManagedCluster --ssh-key-value /path/to/publickey
@@ -806,6 +813,8 @@
           text: az aks create -g MyResourceGroup -n MyManagedCluster --enable-managed-system-pool
         - name: Create a kubernetes cluster with a managed installation of Gateway API CRDs from the standard release channel.
           text: az aks create -g MyResourceGroup -n MyManagedCluster --enable-gateway-api
+        - name: Create a kubernetes cluster with control plane scaling size H4.
+          text: az aks create -g MyResourceGroup -n MyManagedCluster --control-plane-scaling-size H4
         - name: Create an automatic cluster with hosted system components enabled.
           text: az aks create -g MyResourceGroup -n MyManagedCluster --sku automatic --enable-hosted-system
 

src/aks-preview/azext_aks_preview/_params.py

@@ -1261,6 +1261,14 @@ def load_arguments(self, _):
             help="Enable Gateway API based ingress on App Routing via Istio"
         )
         c.argument("enable_hosted_system", action="store_true", is_preview=True)
+        c.argument(
+            "control_plane_scaling_size",
+            options_list=["--control-plane-scaling-size", "--cp-scaling-size"],
+            arg_type=get_enum_type(["H2", "H4", "H8"]),
+            is_preview=True,
+            help="The control plane scaling size. Provides scaled and performance-guaranteed control plane capacity. "
+                 "Available values are 'H2', 'H4', and 'H8'.",
+        )
         c.argument(
             "enable_continuous_control_plane_and_addon_monitor",
             action="store_true",

src/aks-preview/azext_aks_preview/custom.py

@@ -1180,6 +1180,7 @@ def aks_create(
     # app routing istio
     enable_app_routing_istio=False,
     enable_hosted_system=False,
+    control_plane_scaling_size=None,
     # health monitor
     enable_continuous_control_plane_and_addon_monitor=False,
 ):

src/aks-preview/azext_aks_preview/managed_cluster_decorator.py

@@ -3970,6 +3970,13 @@ def get_enable_hosted_system(self) -> bool:
             raise RequiredArgumentMissingError('"--enable-hosted-system" requires "--sku automatic".')
         return enable_hosted_system
 
+    def get_control_plane_scaling_size(self) -> Union[str, None]:
+        """Obtain the value of control_plane_scaling_size.
+
+        :return: str or None
+        """
+        return self.raw_param.get("control_plane_scaling_size")
+
     def get_enable_continuous_control_plane_and_addon_monitor(self) -> bool:
         """Obtain the value of enable_continuous_control_plane_and_addon_monitor.
 
@@ -5069,6 +5076,23 @@ def set_up_bootstrap_profile(self, mc: ManagedCluster) -> ManagedCluster:
 
         return mc
 
+    def set_up_control_plane_scaling_profile(self, mc: ManagedCluster) -> ManagedCluster:
+        """Set up the control plane scaling profile for the ManagedCluster object.
+
+        :return: the ManagedCluster object
+        """
+        self._ensure_mc(mc)
+
+        control_plane_scaling_size = self.context.get_control_plane_scaling_size()
+        if control_plane_scaling_size is not None:
+            mc.control_plane_scaling_profile = (
+                self.models.ManagedClusterControlPlaneScalingProfile(  # pylint: disable=no-member
+                    scaling_size=control_plane_scaling_size,
+                )
+            )
+
+        return mc
+
     def set_up_health_monitor_profile(self, mc: ManagedCluster) -> ManagedCluster:
         """Set up health monitor profile for the ManagedCluster object.
 
@@ -5211,6 +5235,8 @@ def construct_mc_profile_preview(self, bypass_restore_defaults: bool = False) ->
         mc = self.set_up_agentpool_profile_ssh_access(mc)
         # set up bootstrap profile
         mc = self.set_up_bootstrap_profile(mc)
+        # set up control plane scaling profile
+        mc = self.set_up_control_plane_scaling_profile(mc)
         # set up static egress gateway profile
         mc = self.set_up_static_egress_gateway(mc)
         # set up health monitor profile

src/aks-preview/azext_aks_preview/tests/latest/test_aks_commands.py

@@ -9959,6 +9959,53 @@ def test_aks_create_with_windows_gmsa(
             checks=[self.is_empty()],
         )
 
+    @live_only()
+    @AllowLargeResponse()
+    @AKSCustomResourceGroupPreparer(
+        random_name_length=17,
+        name_prefix="clitest",
+        location="centraluseuap",
+    )
+    def test_aks_create_with_control_plane_scaling_profile(
+        self, resource_group, resource_group_location
+    ):
+        # reset the count so in replay mode the random names will start with 0
+        self.test_resources_count = 0
+        # kwargs for string formatting
+        aks_name = self.create_random_name("cliakstest", 16)
+        self.kwargs.update(
+            {
+                "resource_group": resource_group,
+                "name": aks_name,
+                "location": resource_group_location,
+                "resource_type": "Microsoft.ContainerService/ManagedClusters",
+                "ssh_key_value": self.generate_ssh_keys(),
+            }
+        )
+
+        # create with control plane scaling size H4
+        create_cmd = (
+            "aks create --resource-group={resource_group} --name={name} --location={location} "
+            "--network-plugin azure --network-plugin-mode overlay --pod-cidr 10.244.0.0/16 "
+            "--ssh-key-value={ssh_key_value} --node-count 1 "
+            "--control-plane-scaling-size H4"
+        )
+        self.cmd(
+            create_cmd,
+            checks=[
+                self.check("provisioningState", "Succeeded"),
+                self.check("networkProfile.networkPlugin", "azure"),
+                self.check("networkProfile.networkPluginMode", "overlay"),
+                self.check("controlPlaneScalingProfile.scalingSize", "H4"),
+            ],
+        )
+
+        # delete
+        self.cmd(
+            "aks delete -g {resource_group} -n {name} --yes --no-wait",
+            checks=[self.is_empty()],
+        )
+
     @AllowLargeResponse()
     @AKSCustomResourceGroupPreparer(
         random_name_length=17,