potiuk commented on code in PR #33084:
URL: https://github.com/apache/airflow/pull/33084#discussion_r1287411682
##########
airflow/jobs/job.py:
##########
@@ -117,6 +115,15 @@ def __init__(self, executor=None, heartrate=None,
**kwargs):
def executor(self):
return ExecutorLoader.get_default_executor()
+ @cached_property
+ def heartrate(self):
+ if self.job_type == "TriggererJob":
+ return conf.getfloat("triggerer", "JOB_HEARTBEAT_SEC")
+ else:
+ # Heartrate used to be hardcoded to scheduler, so in all other
+ # cases continue to use that value for back compat
+ return conf.getfloat("scheduler", "JOB_HEARTBEAT_SEC")
Review Comment:
Actually this is ok accordig to how `cached_property` works. I recently
stumbled upon that.
The `@cached_property` works different than `@property` -> it actually
creates an attribute that is set the first time the property is used and you
can freely delete or assign the attribute a different value. This is a nice
one, because you can even invalidate the cache by simply deleting the attribute
(which is for example very cool for testing and I am already using it to
invalidat sensitive field list when loading provider configuration
https://github.com/apache/airflow/blob/main/airflow/configuration.py#L1854
It was working like that since 3.8 but it made it into "official part of the
API" in Python 3.9. Similarly as HashMap ordering graduated from being
implementation detail to official behaviour.
--
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: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
