skip to Main Content

Suppose a CI workflow is running, how can I view the full logs?

enter image description here

What I’ve tried

  • I can see the logs if I click on the task immediately after starting a new workflow, for example:

enter image description here

  • But once a workflow has been running I cannot see the full logs
  • I’ve tried clicking on the cog on top right of screen -> View raw logs, but it shows:
2022-07-14T16:29:41.4911629Z Requested labels: ubuntu-latest
2022-07-14T16:29:41.4911713Z Job defined at: username/repo/.github/workflows/main.yml@refs/heads/main
2022-07-14T16:29:41.4911751Z Waiting for a runner to pick up this job...
2022-07-14T16:29:42.2335026Z Job is waiting for a hosted runner to come online.
2022-07-14T16:29:48.5809023Z Job is about to start running on the hosted runner: Hosted Agent (hosted)

But I know there’s much more than that.

  • If I cancel the workflow, it suddenly shows the full logs (I just need to be able to see them without cancelling or restarting a workflow):

enter image description here

2

Answers


  1. This is a bug, and it’s been brought up to their attention long time ago. Still no fix.

    https://github.com/actions/runner/issues/2131

    Login or Signup to reply.
  2. Apparently (which is linked from the issue Marko T mentions), the way the web front end (to Github Actions) works is that it uses a websocket to stream in log lines as they are produced, and issues GETs to bring in lines from terminated steps. The problem comes when a component has hung or is otherwise not producing a lot of output; you don’t get the historical output of it until the step terminates (whether by completing, failing, or being cancelled) and there isn’t much you can do about it. (Possible workaround that I’ve not tested: doing a full fetch of the log might work? It won’t be pretty but at least you’ll be able to see what’s going on.)

    It’s a very frustrating usability issue to say the least, especially as it only really shows up when things are going nastily wrong. Apparently users are supposed to go here to discuss it so that "product managers and engineers will be able to review them" but it looks more like a way of black-holing user requests in somewhere that isn’t actually watched. I could be wrong and maligning people here, but it doesn’t really have that great a look to me as an external user.

    Login or Signup to reply.
Please signup or login to give your own answer.
Back To Top
Search