GitHub Actions reports workflow as failed, even though all steps have succeeded

I am getting a lot of emails about a failed worfklow. Additionally in the overview, the workflow is marked as failed:

When each of the steps run (e.g.: in a PR), everything goes green and succeeds. However, I am still getting mails about the workflow having failed.

1 Like

Thank you for reporting this issue. I have submitted it to the appropriate engineering team for further evaluation. The team will review the feedback. When they give any response, I will update here at the first time. 


Looked at the latest one

it does look like there’s a failure.

I know this is old post, but please let us know if you are still experiencing similar behaviors. If so, a run example would help.

Thank you!

Yes, the issue still persists and is not due to the build errors I have from time to time.

I have the following 5 workflows with a total of 169 jobs:

* Linux
* Docs
* Lint
* Config
* Versions

The one which is always failing (even though it is shown as green and successful) is Versions. Don’t get me wrong, in the end it shows as success, however I receive the email about the workflow as failed.

What makes it even more suspicious is that during the time it is running, the whole workflow fails at some point of time (even though jobs are still running). The running jobs are then shown as failed in the interface with a red X and also the PR pages shows them as failed. Once the jobs finish and succeed, they change their status to succeeded.

Once again, however, I am still receiving the workflow failed email.

Here is one example of a failed, but succeeded workflow:

Thanks, I see it now.

We will create an internal issue tracking this and get this addressed.

Appreciate you following up on this!

Tl;DR I think this is an artifact of having lot of jobs.

just would like to update that there’s a fix that scheduled, in few weeks the issue should be addressed in production.

Thanks for reporting this!

1 Like

@yaananth do you have any update on this?

Sorry, we have it logged and understand the issue, however this is not prioritized yet with the other work. We will prioritize this soon.

This reached deployment stage now :tada:
Hopefully this or next week, changes would be live and hopefully that would address this issue :slight_smile:

This is done now! Please reopen if it’s still happening