You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@airflow.apache.org by "dstandish (via GitHub)" <gi...@apache.org> on 2024/07/08 18:55:44 UTC
[PR] Share data loader to across asyncio boto sessions [airflow]
dstandish opened a new pull request, #40658:
URL: https://github.com/apache/airflow/pull/40658
By default, a botocore session creates and caches an instance of JSONDecoder which consumes a lot of memory. This issue was reported here https://github.com/boto/botocore/issues/3078. In the context of triggers which use boto sessions, this can result in excessive memory usage and as a result reduced capacity on the triggerer. We can reduce memory footprint by sharing the loader instance across the sessions.
Below here is memory utilization shown before and after the fix, with 50 s3 key triggers running:
Before (770MB):
After (370MB):
--
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-unsubscribe@airflow.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] Share data loader to across asyncio boto sessions [airflow]
Posted by "potiuk (via GitHub)" <gi...@apache.org>.
potiuk merged PR #40658:
URL: https://github.com/apache/airflow/pull/40658
--
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-unsubscribe@airflow.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org