GitHub user DTelpuhovsky added a comment to the discussion: Dagbag import is 
taking longer time than an older version, which makes some tasks get timeout 
exceed errors

Sorry for jumping in.
@potiuk I have the similar issue with dynamic dag generation for 1k+ dags from 
json config files.
The performance problems lifecycle looks like this:
- increase amount of json config files --> fixed by splitting to several "dag 
generators"
- parsing problems for scheduling --> fixed by adjustments of scheduler params 
and adding dag processors
- parsing problems on execution with version 2.4.3 --> fixed by experimental 
feature suggested here 
https://airflow.apache.org/docs/apache-airflow/stable/howto/dynamic-dag-generation.html
- performance problems on execution appeared again after an upgrade to 2.7.3 
and persist till 2.10.3

Do you have any idea why does this code stop returning dag_id approx from 
Airflow version 2.4.3?
```
from airflow.utils.dag_parsing_context import get_parsing_context

current_dag_id = get_parsing_context().dag_id

for thing in list_of_things:
    dag_id = f"generated_dag_{thing}"
    if current_dag_id is not None and current_dag_id != dag_id:
        continue  # skip generation of non-selected DAG
```

GitHub link: 
https://github.com/apache/airflow/discussions/44402#discussioncomment-11917207

----
This is an automatically sent email for [email protected].
To unsubscribe, please send an email to: [email protected]

Reply via email to