Skip to content

fix: distinguish configured and live runtime identity - #28

Merged
ShuhaoZhangTony merged 1 commit into
mainfrom
codex/issue-27-runtime-availability
Aug 15, 2026
Merged

fix: distinguish configured and live runtime identity#28
ShuhaoZhangTony merged 1 commit into
mainfrom
codex/issue-27-runtime-availability

Conversation

@ShuhaoZhangTony

Copy link
Copy Markdown
Contributor

What changed

  • distinguish deployment/receipt identity from a verified live serving endpoint
  • prevent configured models from being reported as the model that served a response
  • label Support evidence and health fields with explicit availability
  • extend operational acceptance and regression coverage for live, receipt, timeout, and unknown states

Why

When the inference engine is inactive, deployment receipts still describe the intended model and NPU topology. The application incorrectly presented those facts as live runtime state.

Validation

  • ruff check
  • focused runtime/receipt/operational tests: 80 passed
  • full suite: 728 passed, 28 skipped

Closes #27

@ShuhaoZhangTony
ShuhaoZhangTony merged commit 70c21be into main Aug 15, 2026
5 checks passed
@ShuhaoZhangTony
ShuhaoZhangTony deleted the codex/issue-27-runtime-availability branch August 15, 2026 11:01
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.

[P1][Runtime Identity] Do not claim an inactive inference engine is currently serving

1 participant