Task 2.2: Access workflow artifacts and logs

• AFollow the redirect immediately after each call

  Correct

  The endpoint returns 302 with a download link that expires after one minute.

• BCall the endpoint with Accept: application/zip to get the file inline

  Incorrect

  The endpoint answers with a redirect, not with the archive in the body.

• CUse the jobs endpoint, whose links don't expire

  Incorrect

  Job log downloads also redirect to short-lived links.

• DRequest the logs after the retention period ends

  Incorrect

  After retention ends, the logs are deleted rather than archived.

Download workflow run logs returns 302 Found with a redirect URL to an archive of the run's log files. The link expires after 1 minute, so the client should follow the Location header right away.

• Agh run download -n crash-report from the repository root

  Incorrect

  Without a run ID, the command downloads crash-report artifacts across the repository's runs, not only run 8812.

• Bgh run download 8812

  Incorrect

  Without -n, every artifact of the run is downloaded.

• Cgh run rerun 8812 --debug

  Incorrect

  That starts a new attempt instead of downloading anything.

• Dgh run download 8812 -n crash-report

  Correct

  -n limits the download to the named artifact.

gh run download RUN_ID downloads all artifacts of a run, and -n ARTIFACT_NAME downloads only the named artifact, which is extracted into the current directory when it is the only one.

• ARe-run all jobs so that every job is included in the next attempt's log archive

  Incorrect

  That runs every job again and produces new logs instead of the originals.

• BEnable debug logging and download the archive again

  Incorrect

  Debug logging adds detail to future runs; it doesn't change this archive.

• CDownload the log archive of the earlier run attempt as well

  Correct

  The jobs that weren't re-run are in the archive of the attempt that ran them.

• DRead the logs from the artifacts of the run

  Incorrect

  Logs aren't stored as artifacts unless a workflow uploads them.

When you download the log archive of a workflow that was partially re-run, the archive includes only the jobs that were re-run. For a complete set of logs, download the archives of the previous run attempts that ran the other jobs, which the Latest dropdown lists.

• AThe archive_download_url returned by the REST API

  Incorrect

  That API link needs a token and redirects to a short-lived URL, which a comment can't refresh.

• BThe upload step's artifact-digest output

  Incorrect

  The digest verifies the content; it isn't a link.

• CThe upload step's artifact-url output

  Correct

  The action returns a URL to download that artifact, which works for signed-in users until it expires.

• Dgithub.server_url, the repository and the run ID joined together

  Incorrect

  That points at the run page, which the stem rules out.

actions/upload-artifact returns artifact-id, artifact-url and artifact-digest. artifact-url can link to the artifact from issues or pull requests; users must be signed in, and the URL works until the artifact expires or the run or repository is deleted. The REST download endpoint returns a redirect that expires after a minute.

• AFrom the repository's Releases page

  Incorrect

  Releases hold release assets, not workflow artifacts.

• BFrom the Artifacts section of the run's summary page

  Correct

  Run summaries list the run's artifacts for download.

• CFrom the Packages tab of the repository

  Incorrect

  Packages hold registry packages, not artifacts.

• DFrom the Set up job step of the run log

  Incorrect

  That step records runner setup; it has no downloads.

Open the workflow run from the Actions tab; the run summary's Artifacts section lists the run's artifacts, and clicking one downloads it.

• Agh run view --job JOB_ID --verbose --log

  Incorrect

  --log prints the whole log and --verbose adds the steps; neither keeps only failed steps.

• Bgh run list --limit 1

  Incorrect

  That lists recent runs.

• Cgh run watch

  Incorrect

  That follows a run's progress.

• Dgh run view --job JOB_ID --log-failed

  Correct

  --log-failed filters the log to the failed steps.

gh run view --log prints a job's full log, and --log-failed prints only the logs of failed steps. --verbose adds job steps to the run summary.

• AThe job ID of the uploading job

  Incorrect

  The job isn't used as a name.

• Bartifact

  Correct

  Without a name, upload-artifact uses artifact.

• CThe last folder in path

  Incorrect

  The path doesn't decide the name of a zipped upload.

• DThe run number

  Incorrect

  Run numbers aren't used as names.

GitHub recommends naming each artifact. If no name is provided, the upload-artifact action uses the default name artifact, and download-artifact can download it by that name.

• AEach artifact in its own directory named after it

  Correct

  Downloading all artifacts creates one directory per artifact name.

• BOnly the first artifact is downloaded

  Incorrect

  Without a name, every artifact of the run is downloaded.

• CAll files merged into the workspace root

  Incorrect

  Without further inputs, the artifacts aren't merged into one place.

• DThree zip files that the download step leaves for the job to extract

  Incorrect

  The action extracts zipped artifacts.

download-artifact downloads all artifacts of the run when no name is given, creating a directory for each artifact using its name, and it extracts each zipped artifact.