Skip to content

fix(netops): preserve observed CPU/memory evidence in Ansible facts (PLU-43) - #152

Draft
kayodebristol wants to merge 5 commits into
mainfrom
review/plu-43-health-facts-20261105
Draft

kayodebristol wants to merge 5 commits into
mainfrom
review/plu-43-health-facts-20261105

Conversation

@kayodebristol

Copy link
Copy Markdown
Contributor

M1 review-only patch (PLU-43)

Corrects the actual parsed field keys for Cisco and Nokia CPU/memory collection in the Ansible facts module. Preserves existing scalar output paths only when measurements are valid. Adds explicit OBSERVED / UNKNOWN / UNSUPPORTED source-evidence coverage and avoids silently substituting zero or falling back to Cisco for unknown vendors.

Tests

Adds 14 synthetic pytest cases: actual Cisco/Nokia parser keys, observed zero, malformed/missing/over-range data, invalid memory totals, command exception, unsupported vendors, and partial coverage. No live device I/O.

Limitations

This is not the full PLU-43 fix. The Cisco interface status-table/detailed counter mismatch, role's unsupported gather selectors and flat-vs-nested assertions, Molecule mocks, BGP/OSPF semantics, real vendor command output/transport fixtures and peer review remain unresolved. The PR stays draft; no merge or deployment. Companion documentation/24-case acceptance matrix is draft PR #151, and source health policy must reconcile with stacked #148 → #149 → #150.

No protected branch was written. Do not use this commit as proof of production health or M1 construction readiness.

This branch has not been deployed

No deployments
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