Skip to content

test: validate GPU recording commit-back workflow#6214

Draft
skamenan7 wants to merge 3 commits into
mainfrom
validation/gpu-commitback-20260629
Draft

test: validate GPU recording commit-back workflow#6214
skamenan7 wants to merge 3 commits into
mainfrom
validation/gpu-commitback-20260629

Conversation

@skamenan7

Copy link
Copy Markdown
Collaborator

Temporary draft PR to validate the GPU recording artifact handoff into Commit Recordings. This PR contains only a signed empty commit and should be closed after validation.

skamenan7 and others added 3 commits June 29, 2026 10:41
Signed-off-by: Sumanth Kamenani <skamenan@redhat.com>
Signed-off-by: Sumanth Kamenani <skamenan@redhat.com>
Signed-off-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
Co-Authored-By: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
@github-actions

Copy link
Copy Markdown
Contributor

Recordings committed successfully

Recordings from the integration tests have been committed to this PR.

View commit workflow

@mergify

mergify Bot commented Jul 1, 2026

Copy link
Copy Markdown
Contributor

This pull request has merge conflicts that must be resolved before it can be merged. @skamenan7 please rebase it. https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/working-with-forks/syncing-a-fork

@mergify mergify Bot added the needs-rebase label Jul 1, 2026
meefs pushed a commit to meefs/llama-stack that referenced this pull request Jul 1, 2026
## Summary

Keep the GPU recording setup from failing when the installed prometheus
FastAPI instrumentator package no longer has the exact internal line our
compatibility patch expects.

The patch still applies when the old implementation is present. If the
package has already moved past that shape, the setup action logs that
the patch is not needed and continues.

## Background

PR ogx-ai#6101 added the GPU vLLM recording workflow and was merged first so
the trusted `Commit Recordings` follow-up could be tested from the
default branch workflow code.

After that merge, I used temporary draft PR ogx-ai#6214 to validate the real
commit-back path. That proved the GPU workflow can upload
`recordings-vllm-gpu-gpt-oss-*` artifacts and the trusted `Commit
Recordings` workflow can load the PR metadata and push recording changes
back to the PR branch.

During that validation, the first GPU run failed in vLLM setup because
the metrics route compatibility patch expected an older internal shape
from `prometheus_fastapi_instrumentator`. This PR is the clean follow-up
with only that setup fix. The temporary validation PR should not be
merged.

## Test plan

Validation branch run after this fix:

- Launch GPU EC2 Runner run 28380846166 passed for the base suite.
- compute-pr-info, start-gpu-runner, Setup vLLM GPU, record-mode
integration tests, artifact upload, cleanup, and summary all passed.
- Commit Recordings run 28381881066 downloaded
recordings-vllm-gpu-gpt-oss-base-28380846166-1, loaded PR metadata, and
pushed the recordings commit back to the PR branch.

Local commit checks:

```text
pre-commit hooks run during git commit
check yaml: Passed
check for merge conflicts: Passed
check for missing __init__.py files: Passed
Ensure ogx_api does not import ogx: Passed
Enforce AuthorizedSqlStore usage: Passed
```

## Breaking changes

None.

---------

Signed-off-by: Sumanth Kamenani <skamenan@redhat.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant