diff --git a/src/clusterfuzz/_internal/base/tasks/task_rate_limiting.py b/src/clusterfuzz/_internal/base/tasks/task_rate_limiting.py
index 2101a6b0eb..f12e775f74 100644
--- a/src/clusterfuzz/_internal/base/tasks/task_rate_limiting.py
+++ b/src/clusterfuzz/_internal/base/tasks/task_rate_limiting.py
@@ -20,6 +20,9 @@
 from clusterfuzz._internal.metrics import logs
 from clusterfuzz._internal.system import environment
 
+# TODO(metzman): This needs to be done better in utasks where main may return
+# an error rather than except. This needs to be handled in postprocess.
+
 
 def _get_datetime_now():
   return datetime.datetime.now()
diff --git a/src/clusterfuzz/_internal/bot/tasks/commands.py b/src/clusterfuzz/_internal/bot/tasks/commands.py
index 9db1473618..86f6faabe7 100644
--- a/src/clusterfuzz/_internal/bot/tasks/commands.py
+++ b/src/clusterfuzz/_internal/bot/tasks/commands.py
@@ -203,8 +203,8 @@ def run_command(task_name, task_argument, job_name, uworker_env):
   if should_update_task_status(task_name):
     if not data_handler.update_task_status(task_state_name,
                                            data_types.TaskState.STARTED):
-      logs.info('Another instance of "{}" already '
-                'running, exiting.'.format(task_state_name))
+      logs.info(f'Another instance of "{task_state_name}" already running, '
+                'exiting.')
       raise AlreadyRunningError
 
   result = None
@@ -227,10 +227,10 @@ def run_command(task_name, task_argument, job_name, uworker_env):
     rate_limiter.record_task(success=False)
   except BaseException:
     # On any other exceptions, update state to reflect error and re-raise.
+    rate_limiter.record_task(success=False)
     if should_update_task_status(task_name):
       data_handler.update_task_status(task_state_name,
                                       data_types.TaskState.ERROR)
-      rate_limiter.record_task(success=False)
     raise
   else:
     rate_limiter.record_task(success=True)