Jobs that fail when "Uploading artifacts to coordinator" are classed as "script_failure"
Summary
My organisation hosts its own Gitlab with runners, and we're currently having networking issues that manifest as TLS handshake timeouts when uploading artifacts, so we get errors like this:
ERROR: Uploading artifacts to coordinator... error error=couldn't execute POST against https://git.ligo.org/api/v4/jobs/810306/artifacts?artifact_format=gzip&artifact_type=junit: Post "https://git.ligo.org/api/v4/jobs/810306/artifacts?artifact_format=gzip&artifact_type=junit": net/http: TLS handshake timeout id=810306 token=28ZYtstm
This issue is not to report those errors, but rather to question the fact that these errors are categorised as script_failure
, which seems odd. Should these instead be listed as runner_system_failure
?
Steps to reproduce
Unknown.
Actual behavior
Errors uploading artifacts are classed as script_failure
.
Expected behavior
These failures should be classed as runner_system_failure
.
Relevant logs and/or screenshots
Failing job: https://git.ligo.org/duncanmmacleod/conda/-/jobs/810306
Failing pipeline: https://git.ligo.org/duncanmmacleod/conda/pipelines/128444
Environment description
--
Used GitLab Runner version
Running with gitlab-runner 12.10.1 (ce065b93)