Tag all SkyPilot AWS nodes with "skypilot-user: $USER". (skypilot-org#1515)

* Tag all SkyPilot AWS nodes with "skypilot-user: $USER".

* Use getpass.getuser().

* Don't catch getuser()

Author: concretevitamin

sky/backends/backend_utils.py

@@ -768,7 +768,8 @@ def write_cluster_config(
 
     yaml_path = _get_yaml_path_from_cluster_name(cluster_name)
 
-    # Use a tmp file path to avoid incomplete YAML file being re-used in the future.
+    # Use a tmp file path to avoid incomplete YAML file being re-used in the
+    # future.
     tmp_yaml_path = yaml_path + '.tmp'
     tmp_yaml_path = fill_template(
         cluster_config_template,
@@ -778,6 +779,12 @@ def write_cluster_config(
                 'cluster_name': cluster_name,
                 'num_nodes': num_nodes,
                 'disk_size': to_provision.disk_size,
+                # If the current code is run by controller, propagate the real
+                # calling user which should've been passed in as the
+                # SKYPILOT_USER env var (see spot-controller.yaml.j2).
+                'user': os.environ.get('SKYPILOT_USER', getpass.getuser()),
+
+                # AWS only:
                 # Temporary measure, as deleting per-cluster SGs is too slow.
                 # See https://github.com/skypilot-org/skypilot/pull/742.
                 # Generate the name of the security group we're looking for.

sky/execution.py

@@ -527,6 +527,7 @@ def spot_launch(
                 'disable_logging': env_options.Options.DISABLE_LOGGING.get(),
                 'logging_user_hash': common_utils.get_user_hash(),
                 'retry_until_up': retry_until_up,
+                'user': os.environ.get('USER', None),
             },
             output_prefix=spot.SPOT_CONTROLLER_YAML_PREFIX)
         controller_task = task_lib.Task.from_yaml(yaml_path)

sky/templates/aws-ray.yml.j2

@@ -43,6 +43,11 @@ available_node_types:
           #   SpotOptions:
           #       MaxPrice: MAX_HOURLY_PRICE
       {% endif %}
+      TagSpecifications:
+        - ResourceType: instance
+          Tags:
+            - Key: skypilot-user
+              Value: {{ user }}
 {% if num_nodes > 1 %}
   ray.worker.default:
     min_workers: {{num_nodes - 1}}
@@ -65,6 +70,11 @@ available_node_types:
           #   SpotOptions:
           #       MaxPrice: MAX_HOURLY_PRICE
       {% endif %}
+      TagSpecifications:
+        - ResourceType: instance
+          Tags:
+            - Key: skypilot-user
+              Value: {{ user }}
 {%- endif %}
 
 head_node_type: ray.head.default

sky/templates/spot-controller.yaml.j2

@@ -39,6 +39,7 @@ run: |
 
 envs:
   SKYPILOT_USAGE_USER_ID: {{logging_user_hash}}
+  SKYPILOT_USER: {{user}}
 {% if is_dev %}
   SKYPILOT_DEV: 1
 {% endif %}