fix(evidence): export phase-backed request decision latency - #1158
Conversation
|
Warning Review limit reachedNext included review available in 19 minutes. View limit detailsLimit details: You’ve used the included review currently available. You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository. Review configuration: ⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Advanced Run ID: 📒 Files selected for processing (5)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
…egrated-20260912' into codex/decision-latency-export-20260913
|
Restacked on the updated base branch ( Note for anyone running these locally: both files import 🤖 Addressed by Claude Code |
…y-export-20260913
…y-export-20260913
Scope and lineage
Stacked on #1138 at
1881ef06ed90ee72eb7209434db366c851cd68dc.Current source:
45cc666f9fd52aedf6484b345f30857d7f9d72bf.Current documentation-only head:
0a2626867c0baa6a95ad40f3f00e40008359cca2.Add request-scoped
decision_latency_msto the existing admin export, convertingonly a validated durable acknowledgement. Retain raw nanoseconds, nulls,
authorization, pagination and all routing defaults.
This repairs #1125's measurement intent but does not claim its trace-row contract
fully inherited or authorize closing #1125/#1138. No duplicate clock or estimator.
Root repairs and evidence
Initial field RED: 10 failed, 14 passed.
Independent review found admission values could impersonate missing final
acknowledgement and missing phase selection; both reproduced independently
and repaired at their existing raw projection boundary.
Current focused projection, actual HTTP/restart/export and API contracts:
69 passed in 3.41s, warnings treated as errors, exit 0.
Independent read-only review of this exact source found both findings resolved
and no further specific defect. This is not GitHub approval.
Initial HTTP setup lacked the native extension; the failure is recorded in
the runbook. Current native build uses this source and locked dependencies.
Default full regression: 3,764 passed, 2 skipped in 178.26s, exit 0.
One unclassified request-log fragment remains unresolved; this is not strict
full-suite or clean-diagnostics evidence.
Separately built noneditable core/native wheels passed manifest separation,
isolated import-origin checks and 69 strict contract tests in 24.71s.
Exact artifact hashes are recorded in the validation runbook. Scope is local
macOS ARM64/Python 3.12, not hosted Linux or publication.
Actual GitHub rendering of the complete added documentation section at
cb97ec66was directly inspected in two screenshots, English, 1265 x 712.Wrapped hashes, commands and paragraphs were readable without observed
overlap or clipping. Both production-file and both test-file diffs were also
directly inspected at
0a262686, same viewport/locale, with no observedclipping or overlap. This is not a full product UI or locale audit.
Historical Draft-triggered run 34708601751 skipped all three jobs by policy.
Existing manual workflow run 34708859235
verifies exact
0a262686: Tests and package quality SUCCESS (7m);CodeQL, supply chain and SBOM SUCCESS (4m10s); fuzz remains running.
Hosted full test log reports 3,764 passed, 2 skipped in 158.45s.
This branch-triggered run is not protected merge-result or independent
approval evidence. The inherited Node20 artifact warning is owned by fix(ci): update artifact upload runtime without changing gates #1156.
No observed customer KPI gain, protected merge or deployment is claimed.
Keep Draft while dependent work and acceptance remain outstanding.