This is an automated email from the ASF dual-hosted git repository.
potiuk pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/airflow.git
The following commit(s) were added to refs/heads/main by this push:
new afcb63e806 Fix triggerer HA doc (#32454)
afcb63e806 is described below
commit afcb63e806755f07cb1d435ef050bfaf96f99c4a
Author: Hussein Awala <[email protected]>
AuthorDate: Sun Jul 9 15:49:27 2023 +0200
Fix triggerer HA doc (#32454)
Signed-off-by: Hussein Awala <[email protected]>
---
docs/apache-airflow/authoring-and-scheduling/deferring.rst | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/docs/apache-airflow/authoring-and-scheduling/deferring.rst
b/docs/apache-airflow/authoring-and-scheduling/deferring.rst
index 5d4e6e2634..6b5f124e8f 100644
--- a/docs/apache-airflow/authoring-and-scheduling/deferring.rst
+++ b/docs/apache-airflow/authoring-and-scheduling/deferring.rst
@@ -203,7 +203,7 @@ Triggers are designed from the ground-up to be
highly-available; if you want to
Depending on how much work the triggers are doing, you can fit from hundreds
to tens of thousands of triggers on a single ``triggerer`` host. By default,
every ``triggerer`` will have a capacity of 1000 triggers it will try to run at
once; you can change this with the ``--capacity`` argument. If you have more
triggers trying to run than you have capacity across all of your ``triggerer``
processes, some triggers will be delayed from running until others have
completed.
-Airflow tries to only run triggers in one place at once, and maintains a
heartbeat to all ``triggerers`` that are currently running. If a ``triggerer``
dies, or becomes partitioned from the network where Airflow's database is
running, Airflow will automatically re-schedule triggers that were on that host
to run elsewhere (after waiting 30 seconds for the machine to re-appear).
+Airflow tries to only run triggers in one place at once, and maintains a
heartbeat to all ``triggerers`` that are currently running. If a ``triggerer``
dies, or becomes partitioned from the network where Airflow's database is
running, Airflow will automatically re-schedule triggers that were on that host
to run elsewhere (after waiting (2.1 * ``triggerer.job_heartbeat_sec``) seconds
for the machine to re-appear).
This means it's possible, but unlikely, for triggers to run in multiple places
at once; this is designed into the Trigger contract, however, and entirely
expected. Airflow will de-duplicate events fired when a trigger is running in
multiple places simultaneously, so this process should be transparent to your
Operators.
