Skip to content

fix: cloudwatch logs fetch logic #20814

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
potiuk merged 1 commit into from
Jan 23, 2022
Merged

fix: cloudwatch logs fetch logic #20814

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 2 additions & 4 deletions airflow/providers/amazon/aws/hooks/logs.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -71,9 +71,7 @@ def get_log_events(
| 'ingestionTime' (int): The time in milliseconds the event was ingested.
"""
next_token = None

event_count = 1
while event_count > 0:
while True:
if next_token is not None:
token_arg: Optional[Dict[str, str]] = {'nextToken': next_token}
else:
Expand All @@ -99,7 +97,7 @@ def get_log_events(

yield from events

if 'nextForwardToken' in response:
if next_token != response['nextForwardToken']:
Copy link
Copy Markdown
Contributor

@ferruzzi ferruzzi Jan 11, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Should we use response.get('nextForwardToken', None) to avoid any chance of a KeyError here, or do we know for sure that will always be included in the response?

Copy link
Copy Markdown
Member

@uranusjr uranusjr Jan 12, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Probably OK to just fail if the response format changes. An exception raised here (assuming the hook is used in a DAG) would eventually be caught by the test runner and logged to the task logs, which would clearly signal the problem (the API has an unexpected change in format). If we silently fall back to None, it would be more difficult to debug later if the API changed format unexpected (missing nextForwardToken), or actually returned "nextForwardToken": null (which is expected and signals a valid scenario, indicating we or the user may have logical bugs in the workflow).

Copy link
Copy Markdown
Contributor Author

@ayushchauhan0811 ayushchauhan0811 Jan 14, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@ferruzzi @uranusjr Is there any change required in this PR or is it good to go?

Copy link
Copy Markdown
Contributor

@ferruzzi ferruzzi Jan 14, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

uranusjr covered my concern, but I'm also not a committer so my approval doesn't carry much (any) weight.

next_token = response['nextForwardToken']
else:
return