Missing test results from Github Actions

[tripleee]

Several times recently, visiting Github test failures only revealed

Error: The log was not found. It may have been deleted based on retention settings.

Perhaps something would need to be tweaked in our retention settings? Obviously we would like for the failed test runs to reveal what went wrong.

[tripleee]

For instace, just now, build 30916 failed in this way; though visiting again a little while later displayed the log. Perhaps it's just a timing issue.

[makyen]

I've only ever seen this error on tests which are old (i.e. which are old enough to have had the logs deleted by the retention policy). The default retention policy for public repositories is 90 days. That is also the maximum permitted value. The policy for SmokeDetector is currently 90 days. I'm not aware of it ever being other than 90 days.

It's strange that you're seeing this, as any time when I've specifically been waiting to see the logs for a run (i.e. looking for them immediately upon CI completion, or even watching the results live, I've never seen that response.

However, I have experienced errors resulting from various security/privacy browser settings/extensions interfering with the ability to display the logs, as the logs require access to JavaScript from different servers than the rest of GitHub. Specifically, it requires JavaScript from "actions.githubusercontent.com" and the logs are fetched via AJAX from "pipelines.actions.githubusercontent.com".

If I had to guess, I would have to assume that it's either a timing issue or an artifact of the privacy/security settings in your browser/browser extensions.

Unfortunately, as far as I'm aware, there's nothing we can do about this through settings on GitHub.

[stale]

This issue has been closed because it has had no recent activity. If this is still important, please add another comment and find someone with write permissions to reopen the issue. Thank you for your contributions.

[makyen]

I have a branch with a lot of thread safety changes for multiple issue, including this one. I need the time to thoroughly go through it at least one more time and to run substantial testing, including running it on an active test SD instance for a notable amount of time.