improved comments/docs

plugins/flytekit-papermill/flytekitplugins/papermill/task.py

@@ -86,6 +86,13 @@ class NotebookTask(PythonInstanceTask[T]):
 
         Users can access these notebooks after execution of the task locally or from remote servers.
 
+    .. note:
+
+        By default, print statements in your notebook won't be transmitted to the pod logs/stdout. If you would
+        like to have logs forwarded as the notebook executes, pass the stream_logs argument. Note that notebook
+        logs can be quite verbose, so ensure you are prepared for any downstream log ingestion costs
+        (e.g., cloudwatch)
+
     .. todo:
 
         Implicit extraction of SparkConfiguration from the notebook is not supported.
@@ -139,10 +146,11 @@ def __init__(
 
         # Send the papermill logger to stdout so that it appears in pod logs. Note that papermill doesn't allow
         # injecting a logger, so we cannot redirect logs to the flyte child loggers (e.g., the userspace logger)
-        # and we instead send to stdout directly
+        # and inherit their settings, but we instead must send logs to stdout directly
         if self._stream_logs:
             papermill_logger = logging.getLogger("papermill")
             papermill_logger.addHandler(logging.StreamHandler(sys.stdout))
+            # Papermill leaves the default level of DEBUG. We increase it here.
             papermill_logger.setLevel(logging.INFO)
 
         if not os.path.exists(self._notebook_path):