Context and request
Issue #404 defines baseline/candidate comparison and artifact persistence for performance regression checks, and issue #68 covers attaching module/docs outputs to releases. What is still missing is an explicit requirement that performance benchmark outputs themselves are published and retained as release-associated artifacts (not only transient workflow logs or ad-hoc local files).
Process-PSModule should provide a durable and discoverable artifact path for benchmark evidence so maintainers can audit released versions over time and correlate release decisions with measured performance.
Acceptance criteria
- Performance benchmark outputs (raw benchmark data, normalized comparison input, and rendered comparison report) are produced in a documented artifact layout.
- On release runs, benchmark outputs are attached to the corresponding GitHub Release as named assets.
- Asset naming is deterministic and includes enough identity to map to release version and benchmark scope.
- On non-release runs (PR/default-branch validation), the same benchmark outputs are uploaded as workflow artifacts with configured retention and no dependency on local runner files.
- Workflow output (step summary and/or release notes link section) includes direct links to where benchmark artifacts can be downloaded.
- Documentation explains where artifacts are stored, retention expectations, and how to retrieve historical benchmark evidence for a specific release.
Technical decisions
Implementation plan
Context and request
Issue #404 defines baseline/candidate comparison and artifact persistence for performance regression checks, and issue #68 covers attaching module/docs outputs to releases. What is still missing is an explicit requirement that performance benchmark outputs themselves are published and retained as release-associated artifacts (not only transient workflow logs or ad-hoc local files).
Process-PSModule should provide a durable and discoverable artifact path for benchmark evidence so maintainers can audit released versions over time and correlate release decisions with measured performance.
Acceptance criteria
Technical decisions
Implementation plan