Skip to content

Accept remote telemetry container aliases in preflight#202

Open
fallintoplace wants to merge 1 commit into
NVIDIA:mainfrom
fallintoplace:fix/telemetry-remote-container-aliases
Open

Accept remote telemetry container aliases in preflight#202
fallintoplace wants to merge 1 commit into
NVIDIA:mainfrom
fallintoplace:fix/telemetry-remote-container-aliases

Conversation

@fallintoplace
Copy link
Copy Markdown

Summary

  • treat URI-shaped telemetry container images as valid during preflight
  • allow telemetry container aliases from srtslurm.yaml containers to resolve to remote images
  • add regression coverage for a DCGM exporter alias resolving to an nvcr.io image

Why

Telemetry container fields are resolved through the same containers map as model containers, but telemetry preflight always checked the resolved value as a local file. That rejected remote image aliases even though they are still container images that can be pulled by the runtime container support.

Test plan

  • PYTHONPATH=src uv run --no-project --with pytest --with pyyaml --with jinja2 --with marshmallow --with marshmallow-dataclass --with requests --with rich --with questionary --with ruamel.yaml --with pydantic --with mcp --with matplotlib -- python -m pytest tests/test_validation.py -v
  • uv run --no-project --with ruff -- ruff check src/srtctl/core/validation.py tests/test_validation.py

@fallintoplace fallintoplace marked this pull request as ready for review June 6, 2026 17:19
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.

1 participant