Files
acdl/metrics
Jon Chery b054849a99 docs(P4): metrics catalog + NORTH_STAR integration + trust snapshot + no-humans thesis (REQ-186,191..195,204,210..213)
P4 (Wave 3, docs) — REQ-186, 191, 192, 193, 194, 195, 204, 210, 211, 212, 213

New docs:
- docs/METRICS.md — canonical KPI catalog (grounded/derived/deferred)
- docs/metrics/*.md — 13 per-KPI definition-of-success docs (D-127)
- docs/METRICS_DEFERRED_ROADMAP.md — 8 deferred metrics + hot-path plan + re-eval triggers (REQ-210)
- docs/NO_HUMANS_THESIS.md — thesis defensibility brief (REQ-213)

New tools:
- core/metrics/trust_snapshot.py — 5 trust metrics + chain-integrity verdict + snapshot hash (REQ-211)
- scripts/check_north_star_diff.sh — CI check for NORTH_STAR strategic section changes (REQ-204)

Modified:
- .ciagent/config.json — strategic_direction_file: .ciagent/NORTH_STAR.md (REQ-186)

---ci---
project: acdl
phase: 4
milestone: v1.17
status: execute
---/ci---
2026-08-04 20:05:06 +00:00
..

Nova Metrics Directory

v1.17 — Strategic Direction, Leadership Metrics & Unified Story (D-128)

This directory holds Nova's telemetry/observability artifacts. The metrics layer is Nova-native (D-120): JSONL event log + SQLite cold store + hash-chained Decision Ledger. No Kafka, Prometheus, ClickHouse, or QLDB.

Artifact inventory

Artifact Type Regenerable? Description
events.jsonl Append-only event log No (append-only state) CloudEvents 1.0 envelopes from all emitters (REQ-187)
decision_ledger.db SQLite append-only hash-chain No (append-only state) Decision Ledger: ai.decision.made + attestation.recorded events (REQ-188, D-121)
nova_metrics.db SQLite cold store Yes (regenerate via collector) Normalized fact/dimension tables (REQ-189, P2)
runs/<run_id>.json Per-run manifest Yes (regenerate from events) Run lifecycle: stages, durations, exit, confidence, HITL (REQ-187)
runs/<run_id>/ Durable run artifacts Yes (regenerate from $WORK) Persisted copies of pcr.json, signal.json, event.json, etc. (REQ-187)
lifecycle/<module>-<env>.json Lifecycle report Yes (regenerate from lifecycle runs) Per-module apply/modify/destroy results (REQ-205)
test-results.xml JUnit XML Yes (regenerate via pytest) Test results (REQ-187, P1 addopts)
test-report.json JSON test report Yes (regenerate via pytest) Test results in JSON (REQ-187, P1 addopts)
coverage.json Coverage report Yes (regenerate via pytest) Code coverage (REQ-206, P1 addopts)
powerbi/ PowerBI export Yes (regenerate via powerbi_export) CSV/JSON views for PowerBI ingestion (REQ-190, P3)
TRUST_SNAPSHOT.md Trust snapshot report Yes (regenerate via trust_snapshot) 5 trust metrics + chain-integrity verdict (REQ-211, P4)

Backup + restore

Append-only state (events.jsonl, decision_ledger.db): these are the source of truth. They should be committed to git (events.jsonl) or snapshotted (decision_ledger.db). If lost, they CANNOT be regenerated — the events they captured are gone.

Regenerable artifacts (nova_metrics.db, runs/, lifecycle/, test-results.xml, coverage.json, powerbi/): these are derived from the append-only state + the source signals (REGRESSION_REPORT.json, $WORK/*.json, junit XML). If lost, re-run the collector (core/metrics/collector.py, P2) to rebuild nova_metrics.db, then re-run the PowerBI export (core/metrics/powerbi_export.py, P3) to rebuild powerbi/.

Restore procedure:

  1. Recover events.jsonl + decision_ledger.db from git/snapshot.
  2. python3 core/metrics/collector.py → rebuilds nova_metrics.db.
  3. python3 core/metrics/powerbi_export.py → rebuilds powerbi/.
  4. python3 core/metrics/trust_snapshot.py → rebuilds TRUST_SNAPSHOT.md.

Concurrency model

Single-writer per run: the run manifest writer is the only writer per run. SQLite WAL mode + BEGIN IMMEDIATE prevents concurrent-write corruption on the Decision Ledger (P1 risk mitigation).