[chore]: Gateway replica naming cleanup

src/dstack/_internal/core/backends/aws/compute.py

@@ -57,10 +57,10 @@
 from dstack._internal.core.models.backends.base import BackendType
 from dstack._internal.core.models.common import CoreModel, validate_json_extra_ignore
 from dstack._internal.core.models.gateways import (
-    GatewayComputeConfiguration,
     GatewayLoadBalancerConfiguration,
     GatewayLoadBalancerData,
-    GatewayProvisioningData,
+    GatewayReplicaConfiguration,
+    GatewayReplicaProvisioningData,
 )
 from dstack._internal.core.models.instances import (
     InstanceAvailability,
@@ -524,10 +524,10 @@ def is_suitable_placement_group(
             return False
         return placement_group.configuration.region == instance_offer.region
 
-    def create_gateway(
+    def create_gateway_replica(
         self,
-        configuration: GatewayComputeConfiguration,
-    ) -> GatewayProvisioningData:
+        configuration: GatewayReplicaConfiguration,
+    ) -> GatewayReplicaProvisioningData:
         ec2_resource = self.session.resource("ec2", region_name=configuration.region)
         ec2_client = self.session.client("ec2", region_name=configuration.region)
 
@@ -587,7 +587,7 @@ def create_gateway(
         instance.wait_until_running()
         instance.reload()  # populate instance.public_ip_address
         ip_address = _get_instance_ip(instance, configuration.public_ip)
-        return GatewayProvisioningData(
+        return GatewayReplicaProvisioningData(
             instance_id=instance.instance_id,
             region=configuration.region,
             availability_zone=availability_zone,
@@ -716,10 +716,10 @@ def create_gateway_load_balancer(
             ).model_dump_json(),
         )
 
-    def terminate_gateway(
+    def terminate_gateway_replica(
         self,
         instance_id: str,
-        configuration: GatewayComputeConfiguration,
+        configuration: GatewayReplicaConfiguration,
         backend_data: Optional[str] = None,
     ):
         self.terminate_instance(

src/dstack/_internal/core/backends/azure/compute.py

@@ -61,8 +61,8 @@
 from dstack._internal.core.errors import ComputeError, NoCapacityError
 from dstack._internal.core.models.backends.base import BackendType
 from dstack._internal.core.models.gateways import (
-    GatewayComputeConfiguration,
-    GatewayProvisioningData,
+    GatewayReplicaConfiguration,
+    GatewayReplicaProvisioningData,
 )
 from dstack._internal.core.models.instances import (
     InstanceAvailability,
@@ -233,10 +233,10 @@ def terminate_instance(
             instance_name=instance_id,
         )
 
-    def create_gateway(
+    def create_gateway_replica(
         self,
-        configuration: GatewayComputeConfiguration,
-    ) -> GatewayProvisioningData:
+        configuration: GatewayReplicaConfiguration,
+    ) -> GatewayReplicaProvisioningData:
         if configuration.instance_type is not None:
             # TODO: support instance_type. Requires selecting a VM image to avoid errors like this:
             # > The selected VM size 'Standard_E4s_v6' cannot boot Hypervisor Generation '1'
@@ -306,16 +306,16 @@ def create_gateway(
             resource_group=self.config.resource_group,
             vm=vm,
         )
-        return GatewayProvisioningData(
+        return GatewayReplicaProvisioningData(
             instance_id=vm.name,
             ip_address=public_ip,
             region=configuration.region,
         )
 
-    def terminate_gateway(
+    def terminate_gateway_replica(
         self,
         instance_id: str,
-        configuration: GatewayComputeConfiguration,
+        configuration: GatewayReplicaConfiguration,
         backend_data: Optional[str] = None,
     ):
         self.terminate_instance(

src/dstack/_internal/core/backends/base/compute.py

@@ -29,10 +29,10 @@
 from dstack._internal.core.models.backends.base import BackendType
 from dstack._internal.core.models.compute_groups import ComputeGroup, ComputeGroupProvisioningData
 from dstack._internal.core.models.gateways import (
-    GatewayComputeConfiguration,
     GatewayLoadBalancerConfiguration,
     GatewayLoadBalancerData,
-    GatewayProvisioningData,
+    GatewayReplicaConfiguration,
+    GatewayReplicaProvisioningData,
 )
 from dstack._internal.core.models.instances import (
     InstanceConfiguration,
@@ -556,25 +556,25 @@ class ComputeWithGatewaySupport(ABC):
     """
 
     @abstractmethod
-    def create_gateway(
+    def create_gateway_replica(
         self,
-        configuration: GatewayComputeConfiguration,
-    ) -> GatewayProvisioningData:
+        configuration: GatewayReplicaConfiguration,
+    ) -> GatewayReplicaProvisioningData:
         """
-        Creates a gateway instance.
+        Creates a gateway replica instance.
         """
         pass
 
     @abstractmethod
-    def terminate_gateway(
+    def terminate_gateway_replica(
         self,
         instance_id: str,
-        configuration: GatewayComputeConfiguration,
+        configuration: GatewayReplicaConfiguration,
         backend_data: Optional[str] = None,
     ):
         """
-        Terminates a gateway instance. Generally, it passes the call to `terminate_instance()`,
-        but may perform additional work such as deleting a load balancer when a gateway has one.
+        Terminates a gateway replica instance. Generally, it passes the call to
+        `terminate_instance()`, but may perform additional work if necessary.
         """
         pass
 
@@ -631,7 +631,7 @@ def deregister_gateway_replica_from_load_balancer(
 class ComputeWithPrivateGatewaySupport:
     """
     Must be subclassed to support private gateways.
-    `create_gateway()` must be able to create private gateways.
+    `create_gateway_replica()` must be able to create private gateways.
     """
 
     pass
@@ -751,15 +751,15 @@ def generate_unique_instance_name_for_job(
 
 
 def generate_unique_gateway_instance_name(
-    gateway_compute_configuration: GatewayComputeConfiguration,
+    gateway_replica_configuration: GatewayReplicaConfiguration,
     max_length: int = _DEFAULT_MAX_RESOURCE_NAME_LEN,
 ) -> str:
     """
     Generates a unique gateway instance name valid across all backends.
     """
     return generate_unique_backend_name(
-        resource_name=gateway_compute_configuration.instance_name,
-        project_name=gateway_compute_configuration.project_name,
+        resource_name=gateway_replica_configuration.instance_name,
+        project_name=gateway_replica_configuration.project_name,
         max_length=max_length,
     )
 

src/dstack/_internal/core/backends/gcp/compute.py

@@ -57,8 +57,8 @@
 from dstack._internal.core.models.backends.base import BackendType
 from dstack._internal.core.models.common import CoreModel, validate_extra_ignore
 from dstack._internal.core.models.gateways import (
-    GatewayComputeConfiguration,
-    GatewayProvisioningData,
+    GatewayReplicaConfiguration,
+    GatewayReplicaProvisioningData,
 )
 from dstack._internal.core.models.instances import (
     InstanceAvailability,
@@ -561,10 +561,10 @@ def are_placement_groups_compatible_with_reservations(self, backend_type: Backen
         # Instead, we use the placement policy defined in reservation settings.
         return False
 
-    def create_gateway(
+    def create_gateway_replica(
         self,
-        configuration: GatewayComputeConfiguration,
-    ) -> GatewayProvisioningData:
+        configuration: GatewayReplicaConfiguration,
+    ) -> GatewayReplicaProvisioningData:
         if self.config.vpc_project_id is None:
             gcp_resources.create_gateway_firewall_rules(
                 firewalls_client=self.firewalls_client,
@@ -628,18 +628,18 @@ def create_gateway(
         instance = self.instances_client.get(
             project=self.config.project_id, zone=zone, instance=instance_name
         )
-        return GatewayProvisioningData(
+        return GatewayReplicaProvisioningData(
             instance_id=instance_name,
             region=configuration.region,  # used for instance termination
             availability_zone=zone,
             ip_address=_get_instance_ip(instance, configuration.public_ip),
             backend_data=json.dumps({"zone": zone}),
         )
 
-    def terminate_gateway(
+    def terminate_gateway_replica(
         self,
         instance_id: str,
-        configuration: GatewayComputeConfiguration,
+        configuration: GatewayReplicaConfiguration,
         backend_data: Optional[str] = None,
     ):
         self.terminate_instance(

src/dstack/_internal/core/backends/kubernetes/compute.py

@@ -79,8 +79,8 @@
 from dstack._internal.core.models.backends.base import BackendType
 from dstack._internal.core.models.common import CoreModel, validate_json_extra_ignore
 from dstack._internal.core.models.gateways import (
-    GatewayComputeConfiguration,
-    GatewayProvisioningData,
+    GatewayReplicaConfiguration,
+    GatewayReplicaProvisioningData,
 )
 from dstack._internal.core.models.instances import (
     InstanceOfferWithAvailability,
@@ -462,10 +462,10 @@ def terminate_instance(
         if not all(deleted):
             raise ComputeError("Not all objects were deleted, check logs")
 
-    def create_gateway(
+    def create_gateway_replica(
         self,
-        configuration: GatewayComputeConfiguration,
-    ) -> GatewayProvisioningData:
+        configuration: GatewayReplicaConfiguration,
+    ) -> GatewayReplicaProvisioningData:
         cluster = self.region_cluster_map.get(configuration.region)
         if cluster is None:
             raise ComputeError(f"Unknown region: {configuration.region!r}")
@@ -577,19 +577,19 @@ def create_gateway(
         if address is None:
             self.terminate_instance(instance_name, region=configuration.region)
             raise ComputeError(
-                "Failed to get gateway hostname. "
+                "Failed to get gateway replica hostname. "
                 "Ensure the Kubernetes cluster supports Load Balancer services."
             )
-        return GatewayProvisioningData(
+        return GatewayReplicaProvisioningData(
             instance_id=instance_name,
             ip_address=address,
             region=cluster.region,
         )
 
-    def terminate_gateway(
+    def terminate_gateway_replica(
         self,
         instance_id: str,
-        configuration: GatewayComputeConfiguration,
+        configuration: GatewayReplicaConfiguration,
         backend_data: Optional[str] = None,
     ):
         region = configuration.region
@@ -600,7 +600,7 @@ def terminate_gateway(
             if cluster is not None:
                 logger.warning(
                     (
-                        "Terminating gateway %s in unknown region %s."
+                        "Terminating gateway replica %s in unknown region %s."
                         " Assuming it was created before multi-cluster support was added"
                         " and is located in cluster %s"
                     ),

src/dstack/_internal/core/models/gateways.py

@@ -187,7 +187,7 @@ class ApplyGatewayPlanInput(CoreModel):
     ] = None
 
 
-class GatewayComputeConfiguration(CoreModel):
+class GatewayReplicaConfiguration(CoreModel):
     project_name: str
     instance_name: str
     backend: BackendType
@@ -199,7 +199,7 @@ class GatewayComputeConfiguration(CoreModel):
     tags: Optional[Dict[str, str]] = None
 
 
-class GatewayProvisioningData(CoreModel):
+class GatewayReplicaProvisioningData(CoreModel):
     instance_id: str
     # TODO: rename `ip_address`; Kubernetes uses domain names here.
     ip_address: str