🎨Autoscaling: warm buffers: create at minimum 1 machine at a time instead of asking directly for the required number (#8262)

Author: sanderegg

services/autoscaling/src/simcore_service_autoscaling/modules/cluster_scaling/_warm_buffer_machines_pool_core.py

@@ -314,7 +314,7 @@ async def _add_remove_buffer_instances(
                 subnet_id=app_settings.AUTOSCALING_EC2_INSTANCES.EC2_INSTANCES_SUBNET_ID,
                 iam_instance_profile=app_settings.AUTOSCALING_EC2_INSTANCES.EC2_INSTANCES_ATTACHED_IAM_PROFILE,
             ),
-            min_number_of_instances=num_to_start,
+            min_number_of_instances=1,  # NOTE: we want at least 1
             number_of_instances=num_to_start,
             max_total_number_of_instances=app_settings.AUTOSCALING_EC2_INSTANCES.EC2_INSTANCES_MAX_INSTANCES,
         )

services/autoscaling/tests/unit/test_modules_cluster_scaling_warm_buffer_machine_core.py

@@ -268,6 +268,7 @@ async def _assert_wait_for_ssm_command_to_finish() -> None:
 
 
 async def test_monitor_buffer_machines(
+    patch_ec2_client_launch_instances_min_number_of_instances: None,
     minimal_configuration: None,
     ec2_client: EC2Client,
     buffer_count: int,
@@ -374,6 +375,7 @@ async def test_monitor_buffer_machines_terminates_supernumerary_instances(
 
 
 async def test_monitor_buffer_machines_terminates_instances_with_incorrect_pre_pulled_images(
+    patch_ec2_client_launch_instances_min_number_of_instances: None,
     minimal_configuration: None,
     ec2_client: EC2Client,
     buffer_count: int,
@@ -457,6 +459,7 @@ def unneeded_instance_type(
     ],
 )
 async def test_monitor_buffer_machines_terminates_unneeded_pool(
+    patch_ec2_client_launch_instances_min_number_of_instances: None,
     minimal_configuration: None,
     ec2_client: EC2Client,
     buffer_count: int,