Ensure done before checking success to help troubleshoot timeout issues #4833
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
I think this can help with #4624.
It's a little delicate here, but in the case of a "bad" error code that we don't expect a retry for, we try and verify that it both failed and that the attempt was just 1. In the flaky example run linked in #4624, the attempt count was 0, which strongly suggests a race condition. That is, I think we were checking the count before the operation has completed.
I think what happened is that the
.isSuccess()
call fails both when the underlying call fails but also when the call has not completed. By putting an explicit.isDone()
check after the.join()
we can then tell if/when this happens again and then bump up the wait time. I suspect we just hit a cloud glitch and got stalled/lags, as happens sometimes. With this change, we should know.