Clario / Project Health Dashboard

GitHub sync audit & engineering metrics · repo ashispoddar/ezpay · CLARIO project #1

Window: Apr 12 → Jun 15, 2026 · ~9 weeks

Generated 2026-06-15

Issues

20 closed · 14 open

PRs merged

1 open · 1 abandoned

124

Commits

since Apr 12

1.8h

Avg merge latency

max 47h

+81.6k

LOC added

−1.7k removed

Sync gaps

1 high · 2 med · 3 low

① Issue / Project Sync Audit

ID	Sev	Finding	Fix
F1	HIGH	Epic #59 "M8 — Pilot Hardening" still Open / board In progress, but all 5 children (#92–96) closed & PRD-2 done. Checkboxes unticked.	Tick, close, → Done
F2	MED	3 closed bugs #28, #35, #36 never added to CLARIO board — invisible to tracking.	item-add → Done
F3	MED	PR #68 (Flyway runbook + seeder) open & non-draft since 06-13.	Merge or close
F4	LOW	PRD-1 (Card Issuing) has zero issues/board items despite being done — no traceability.	Backfill closed epic
F5	LOW	Milestone field unused on all 34 issues.	Optional milestones
F6	LOW	#100/#84/#65 carry only generic enhancement label (PRD-3 is the model).	Add epic/area labels

30 of 31 tracked items match issue state — board is healthy; gaps are concentrated and quick to fix. Ordered remediation (commands in the markdown report): close #59 → adopt orphan bugs → resolve PR #68 → backfill PRD-1 → hygiene.

CLARIO BOARD STATUS · 31 items

Done · 17 Backlog · 13 In progress · 1

Backlog after cleanup: PRD-3 Card Auth (#70–76), Analytics (#100), infra #32/#39/#60/#65/#84.

② Engineering Metrics — last ~9 weeks

COMMIT CADENCE BY WEEK

Two surges: mid-Apr scaffold (W15) and the Java strangler migration + expense module (W21–24). Gaps = design/planning periods.

COMMIT TYPE MIX (conventional prefix)

feat27

fix16

docs3

chore3

test2

refactor1

ci1

~53% feat / ~31% fix — feature-forward, fits pre-pilot. Thin test/refactor prefixes understate reality: tests ship inside feat/fix commits (21 Java + 20 e2e specs).

CODE CHURN BY AREA (+added)

core-service (Java)48k

frontend/src29k

fe lockfile8k

frontend/e2e2.5k

parity-harness1.3k

infra/lib (CDK)0.9k

Java core-service is net-additive (+48k, ~0 deletions) = strangler built alongside, parity-tested. Frontend holds the only real deletion volume (−2.6k) = genuine iteration.

REPO COMPOSITION & THROUGHPUT

424

Java files

106

TS/TSX

test files

Avg PR size	~1,236 add / 26 del / 14 files
PRs merged by month	Apr 1 · May 19 · Jun 46
Java tests / e2e specs	21 / 20
Test : Java-file ratio	~1 : 20

③ Engineering Observations

Healthy review loop, thin safety net for size. 1.8h merge latency is excellent, but +1.2k-LOC avg PRs with ~1 test file per 20 Java files mean test density trails feature density. Raise coverage expectations on the PRD-3 money-movement path (auth/settle/disputes).
Migration discipline is strong. Net-additive Java service + dedicated parity-harness + per-slice squash PRs (A0–A4) is a textbook strangler. Keep the parity gate in CI until the Node catch-all is retired.
Two correctness deltas remain open from the migration (#39 JSON number format, #60 Textract parse) — small cross-language drift; clear before A5.
Process consistency is uneven across PRDs. PRD-3's epic + prdN-* label + linked sub-issues is the model; PRD-1 (untracked) and analytics/infra items fall short. Standardize on it for trustworthy velocity reporting.

Self-contained · zero external dependencies · host on any static target (S3 / Netlify / GitHub Pages) Source: docs/reports/2026-06-15-gh-sync-and-eng-metrics.md