G2 S8 follow-up: probe-count fact fix, narration node-label correction, provenance/currency tails#163
Merged
Merged
Conversation
…rration, done-status notes - README, architecture.md, RELEASE_STATUS, clickhouse-cutover-plan, ADR 0006: the live '32/32 adversarial probes' claim was stale — the checked-in governance/verify_live.sh asserts 29 probes and the 2026-07-03 rerun passed 29/29 (that transcript already flags the old 32). Release blurbs now say 'every live adversarial probe green' (no brittle count); docs citing a verification record now say 29/29 and point at the 07-03 transcript. CHANGELOG and dated docs/perf transcripts left untouched. - demo_voiced.narration.txt: the S5 rewrite claimed the three demo nodes are labeled branch=msk/dxb/ala with nodeSelector pinning by that label; per demo_evidence.md and the transcript, all three nodes are branch=msk and placement is pinned by workload=postgres|clickhouse — per-branch labels are the production topology. Fixed before S6 records TTS audio from this text. - generator-spec.md §6: 'the current reference is grocery' → past tense; the kitchen-appliance reference replaced it in S2b. - domain.md §5.4: RUB demo-value repin row 'Planned' → Done (S0/S1). Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
…nts; repin residual USD tails to RUB - The three remaining comments naming the retired benchmark dataset (two DDL-guard test headers, vault_rows.py docstring) now describe the same provenance without the literal name, pointing at CHANGELOG.md — consistent with the removal policy and with domain.md's own wording. - Residual USD tails outside the accepted M1 enum scope: Grafana merch-agent-journey revenue panels currencyUSD → currencyRUB, both freshness benchmark event builders Currency.USD → Currency.RUB, and the backup workflow's synthetic restore-check row 'USD' → 'RUB'. No test asserts these values; Currency.RUB is the documented house default. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
DORA Metrics
MTTR note: No failed mainline CI runs in the selected window. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Follow-up cleanup after the S8 control re-audit (plan: new_plen_05_07_26.md). Textual/factual only — no logic changes.
Item 1 — "32/32 probes" → 29 (verified)
warehouse/agentflow/dv2/governance/verify_live.shasserts 29 probes (grep -c 'probe "'= 29); the 2026-07-03 perf transcript passed 29/29 and itself flags the stale "32". Live-facing docs fixed:README.md,docs/dv2-multi-branch/RELEASE_STATUS.md(release blurbs) — reworded to "every live adversarial probe green" (no brittle count).docs/architecture.md,docs/clickhouse-cutover-plan.md,docs/decisions/0006— now say 29/29 and citevault-pii-governance-verify-2026-07-03.md.CHANGELOG.mdand the dateddocs/perf/transcripts intentionally untouched (point-in-time records).Item 2 — generator-spec.md §6 present-tense grocery reference
"The current reference is grocery … it is replaced wholesale" → past tense: the grocery reference was replaced with the kitchen-appliance reference (S2b); same for the
tnved.pybullet.Item 3 — demo narration factual error (pre-TTS fix)
demo_voiced.narration.txtclaimed the three demo nodes are labeledbranch=msk/dxb/alawith nodeSelector pinning by that label. Reality (demo_evidence.md §1–2, demo_transcript.txt kubectl output): all three nodes arebranch=msk; placement is pinned viaworkload=postgres|clickhouse; per-branch labels are the production topology. Narration corrected, voice/pacing kept.demo_transcript.txtneeded no change — its kubectl output was already correct.Item 4 — domain.md §5.4 stale "Planned"
RUB demo-value repin row → Done (S0/S1), matching the neighboring row's style.
Item 5 — retired-dataset name in 3 provenance comments (judgment call: reworded)
The two DDL-guard test headers and the
vault_rows.pydocstring described removal history but still spelled the dataset name — ironically the same comments explain why the tests avoid spelling its prefix. Reworded to "retired benchmark dataset / legacy external-dataset loader" with a CHANGELOG.md pointer; informational value (why the guard/models exist) preserved. Name survives only in CHANGELOG history.Item 6 — residual USD tails (scope calls)
Done (all verified assertion-free):
merch-agent-journey.json: 2×currencyUSD→currencyRUB(valid Grafana unit; JSON re-validated).scripts/benchmark_freshness.py+_realpath.py:Currency.USD→Currency.RUB(enum's documented house default; the only covering test checks pipeline-validator pass, not currency)..github/workflows/backup.yml: synthetic restore-check row"USD"→"RUB"(sole occurrence, nothing asserts on it).Left alone deliberately:
test_search.py,test_event_schemas.py, SDK tests) — valid under the accepted M1 scope; renaming catalog fixtures risks touching search-match semantics for zero factual gain.notebooks/02-nl-query.ipynb'electronics' — it is inside a recorded execution output cell; hand-editing would fabricate an execution result. Right fix is a notebook re-run against the current catalog (candidate for a later step).Verification
ruff check src/ tests/ scripts/— clean;ruff format --check— 361 files already formatted;mypy src/ --ignore-missing-imports— no issues in 124 files.pytest tests/unit/test_dv2_business_vault_ddl.py test_dv2_postgres_ddl.py test_benchmark_freshness.py test_dv2_postgres_ingestion.py— 95 passed (one transient temp-dir flake on first run, not reproducible, passes in isolation and on rerun).-Xmx512m/X509.🤖 Generated with Claude Code