aeroyorch opened a new pull request, #73044:
URL: https://github.com/apache/airflow/pull/73044
Deferrable `KubernetesPodOperator` can exhaust triggerer memory when
`get_logs` and
`logging_interval` are set.
`read_logs` kept the raw bytes, the decoded string and the split line list
in memory at once, then
queued them on the default thread executor, whose queue is unbounded. So
peak memory scaled with the
log window, about 2.5 times its size.
The window is not bounded either. The trigger only keeps the timestamp of
the last log it read in
memory, so after a restart or a reassignment it reads the pod log from the
beginning. On a long
running verbose pod that pins hundreds of MB per trigger, the triggerer is
`OOMKilled`, and the next
restart reads from the beginning again.
This PR fixes the memory side. **The lost timestamp needs its own fix**.
Now the response is streamed line by line, so peak memory is one chunk
whatever the window size.
The loop stays responsive as in #69661. Each chunk yields explicitly,
because the reader only awaits
when its buffer drains, which can be thousands of lines apart.
With 4 / 32 / 128 MiB windows, peak memory drops from 10 / 82 / 326 MiB to
0.5 MiB, and the longest
event loop stall no longer grows with the window. CPU cost is about the same.
**Two behaviour changes to flag to reviewers**: a failed read is retried at
the next `logging_interval` instead of
immediately, and lines are split on `\n` (plus CRLF) rather than
`str.splitlines()`.
<!-- SPDX-License-Identifier: Apache-2.0
https://www.apache.org/licenses/LICENSE-2.0 -->
<!--
Thank you for contributing!
Please provide above a brief description of the changes made in this pull
request.
Write a good git commit message following this guide:
https://chris.beams.io/posts/git-commit/
Please make sure that your code changes are covered with tests.
And in case of new features or big changes remember to adjust the
documentation.
For user-facing UI changes, please attach before/after screenshots (or a
short
screen recording) so reviewers can assess the visual impact.
Feel free to ping (in general) for the review if you do not see reaction for
a few days
(72 Hours is the minimum reaction time you can expect from volunteers) - we
sometimes miss notifications.
In case of an existing issue, reference it using one of the following:
* closes: #ISSUE
* related: #ISSUE
-->
---
##### Was generative AI tooling used to co-author this PR?
<!--
If generative AI tooling has been used in the process of authoring this PR,
please
change below checkbox to `[X]` followed by the name of the tool, uncomment
the "Generated-by".
-->
- [X] Yes (please specify the tool below)
Claude Code (Opus 5) for the streaming implementation, tests and benchmark.
<!--
Generated-by: Claude Code (Opus 5) following [the
guidelines](https://github.com/apache/airflow/blob/main/contributing-docs/05_pull_requests.rst#gen-ai-assisted-contributions)
-->
---
* Read the **[Pull Request
Guidelines](https://github.com/apache/airflow/blob/main/contributing-docs/05_pull_requests.rst#pull-request-guidelines)**
for more information. Note: commit author/co-author name and email in commits
become permanently public when merged.
* For fundamental code changes, an Airflow Improvement Proposal
([AIP](https://cwiki.apache.org/confluence/display/AIRFLOW/Airflow+Improvement+Proposals))
is needed.
* When adding dependency, check compliance with the [ASF 3rd Party License
Policy](https://www.apache.org/legal/resolved.html#category-x).
* For significant user-facing changes create newsfragment:
`{pr_number}.significant.rst`, in
[airflow-core/newsfragments](https://github.com/apache/airflow/tree/main/airflow-core/newsfragments).
You can add this file in a follow-up commit after the PR is created so you
know the PR number.
--
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]
