Skip to content

fix(ci): surface failures in the vLLM nightly workflow - #1488

Open
voropaevv wants to merge 1 commit into
huggingface:mainfrom
voropaevv:codex/report-vllm-canary-failure
Open

voropaevv wants to merge 1 commit into
huggingface:mainfrom
voropaevv:codex/report-vllm-canary-failure

Conversation

@voropaevv

Copy link
Copy Markdown

The scheduled vLLM workflow reports success even when its only job fails. The November 9 run is one example.

Remove the job-level continue-on-error so GitHub Actions reports a failed test job as a failed run. The workflow runs on its weekly schedule and by manual dispatch.

Validated with actionlint, the full pre-commit suite, and an independent review of the YAML change. The GPU job was not dispatched.

Fixes #1486.

Codex assisted with implementation and verification.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

[BUG] vllm_main_tests.yaml: continue-on-error hides a canary 23 scheduled runs deep in failure

1 participant