potiuk commented on code in PR #30361:
URL: https://github.com/apache/airflow/pull/30361#discussion_r1157151324


##########
airflow/executors/base_executor.py:
##########
@@ -216,6 +215,8 @@ def heartbeat(self) -> None:
         self.log.debug("%s in queue", num_queued_tasks)
         self.log.debug("%s open slots", open_slots)
 
+        from airflow.stats import Stats

Review Comment:
   I think those measurments for stats come from running it without "from 
airflow.configuration import conf".  This is one of the most complex imports, 
and is about the only thing that seem to have measureable impect when importing 
stats module. 
   
   If you look up the file it is already imported  up there, conf is imported 
up there, so I think @uranusjr is right about this one, Same with the 
"Exceptions"
   
   Also why the measurement can be biased (for exceptions case) is that when we 
are importing any airflow package, we  also import `airflow.__init__` and this 
one is also rather slow-ish  with configuration and settings import. I tihnk 
this might have (If I am not mistaken) significant impct on the times measured 
for the imports done in isolation.
   
   My tests for one have shown that once `airflow.configuration.conf` is 
imported, stats and exceptions import instanteneously.
   



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscr...@airflow.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org

Reply via email to