Logs won't open.

When I am looking at running actions, most of the time, I am unable to open the logs. Sometimes it opens, though.

The URL I am on is https://github.com/:owner:/:repo:/runs/:runID:

I try click on an action to expand it and show the real-time logs, but it doesn’t open usually. Sometimes if I wait long enough, it will open.

Is this a problem with github actions or is there something I can do on my end to make it work?
Is this the proper place to report bugs with github actions?

Hi @taylorgoolsby ,

The URL cannot be accessed, is it private or deleted? Could you please provide it again? Or please provide log screenshot and yaml file content to help for investigation.

Please try to refesh the webpage and get the real-time logs again, it could take some seconds to download the logs.

You’re welcome to share your query here,  more detail policy please check the link here.

Thanks.

Frequently the Actions pipeline will just get stuck and apparently do nothing for 6 hours, after which it will be terminated.

What I do is that I’m trying to diagnose the problem by checking the build logs, whether the freeze originates from the build itself or is a Github Actions glitch.

However, I’m even unable to see the logs.

I’m really dissatisfied with the diagnosing capabilities when something goes wrong, and by workflows getting stuck randomly like that.

That logs might not load properly while the workflow is running is unfortunately a known issue, see this thread for example: How to see the full log while a workflow is in progress?

I’ve had hangs in builds too, the means I used to debug them were, simplified:

  • Check logs after cancelling the workflow. I haven’t seen that fail at least. :sweat_smile:
  • Upload possibly helpful log files created by the build in a step after the hanging one, with if: failure() || cancelled() (here’s an example in one of my workflows).

A catch with the upload is that I’m not sure if it works if the job gets cancelled by timeout, you might have to cancel manually, or try setting jobs.<job_id>.steps.timeout-minutes on the hanging step (sounds like it should work, but I haven’t tried yet).