Skip to content

[o11y] Drop warning for exceptions reported with non-exception outcome - #7665

Open
fhanau wants to merge 1 commit into
mainfrom
felix/100826-drop-warning
Open

fhanau wants to merge 1 commit into
mainfrom
felix/100826-drop-warning

Conversation

@fhanau

@fhanau fhanau commented Oct 8, 2026 •

Copy link
Copy Markdown
Contributor

This is actually the expected behavior in several cases and should not be a cause for concern:

  • Non-exception outcomes other than OK may describe the ground truth more accurately than "exception", e.g. when exceededCpu is reported for an "exceeded CPU time limit" exception.
  • A background task failure may lead to a canceled outcome – we should report the exception so the developer can understand what caused the cancellation.
  • For some disconnects, we report OK outcomes to avoid noise but may still report exceptions caused by the disconnect like "Network connection lost" to indicate that the request actually disconnected.
  • We may get exceptions from waitUntil tasks that finish running after a fetch request has already returned its response successfully – in that case it is only right to report an OK outcome since the fetch succeeded while still reporting the exception so that the developer is aware of the task failure.
  • I was able to identify one case where this is genuinely an issue: This is being addressed in [o11y] Make sure that exceptions are reported for the right request in actors #7578.

This warning helped us identify some real-world issues, but at the end of the day the assumption that exceptions should only be reported when there's an exception outcome don't really reflect how worker requests behave in practice.

#7578 should land before this PR. Internal PR !15327 is also related but may actually increase the volume of this warning since it affects one of the benign cases, so it's probably better to land this first.

This is actually the expected behavior in several cases and should not be a
cause for concern:
- Non-exception outcomes other than OK may describe the ground truth more
  accurately than "exception", e.g. when exceededCpu is reported for an
  "exceeded CPU time limit" exception.
- A background task failure may lead to a canceled outcome – we should report
  the exception so the developer can understand what caused the cancellation.
- For some disconnects, we report OK outcomes to avoid noise but may still
  report exceptions caused by the disconnect like "Network connection lost" to
  indicate that the request actually disconnected.
- We may get exceptions from waitUntil tasks that finish running after a fetch
  request has already returned its response successfully – in that case it is
  only right to report an OK outcome since the fetch succeeded while still
  reporting the exception so that the developer is aware of the task failure.
- I was able to identify one case where this is genuinely an issue: This is
  being addressed in #7578.
@fhanau
fhanau requested review from a team as code owners October 8, 2026 14:49
@ask-bonk

ask-bonk Bot commented Oct 8, 2026 •

Copy link
Copy Markdown
Contributor

LGTM!


Reviewed commit: 02970b7d · github run

@fhanau
fhanau requested a review from mar-cf October 19, 2026 01:34
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant