NHTSA Health

Operator dashboard for the NHTSA harvesting subsystem. Read-only Phase 0 scaffold; live data wires in Phases 1–5.

Data Types in Catalog

21

9 flat files · 8 JSON APIs · 4 S3 PDF listings

Rolling Window

2017–2026

10 model years (WINDOW_YEARS=10)

Open Root Causes

5

Of 5 tracked Stage 1.7 root causes

Last Audit

Audit freshness pending Phase 1

Root Causes (Stage 1.7)

RC-1

URL Drift (stale flat-file URLs)

6 confirmed drift cases as of Stage 1.5. Decodo HEAD/GET parity pending Phase 2.

Pending data
Remediation: Phase 2
RC-2

Akamai Stale ETag (HEAD blocked)

Datacenter IP block confirmed; residential proxy bypass validated. Wire-up pending Phase 2.

Pending data
Remediation: Phase 2
RC-3

Circuit Breaker Stuck Open

Per-year breakers + resumable runs pending Phase 4.

Pending data
Remediation: Phase 4
RC-4

Distributor Never Run

ekis-zero/source/nhtsa/ + ekis-one/source/nhtsa/ are empty. Distributor cron pending Phase 3.

Pending data
Remediation: Phase 3
RC-5

Operator Visibility Gap

This dashboard is the Phase 0 fix. Live data wires in Phases 1-5.

Pending data
Remediation: Phase 0

10-Year Rolling Window Enforcement

Source-of-truth window from @ekis/config. Vehicles outside this window are out of scope for harvest, ingest, and MILA retrieval.

MY
2017
MY
2018
MY
2019
MY
2020
MY
2021
MY
2022
MY
2023
MY
2024
MY
2025
MY
2026
Per-year coverage gauge wires in Phase 4 (post-backfill).

URL Drift

Pending data

Decodo HEAD/GET parity sweep + URL canonicalization — pending Phase 2.

Wires in Phase 2

Akamai ETag Freshness

Pending data

Residential-proxy ETag refresh job — pending Phase 2.

Wires in Phase 2

Source Library Drift

Pending data

R2 prefix counts across the three tiers — pending Phase 3 distributor unfreeze.

Wires in Phase 3

Per-Data-Type Circuit Breakers

Pending data

IR-6 circuit breaker state per data type — pending Phase 4.

Wires in Phase 4

Audit Truncation

Pending data

Detect bimodal audit sizes (4039 byte truncation vs ~19KB full) — pending Phase 1.

Wires in Phase 1

Live Pipeline Heartbeat

Pending data

PL-3 real-time progress feed for active pipeline runs — pending Phase 1.

Wires in Phase 1

Data Type Contract Matrix (Stage 1.8)

ChannelData TypeContractRows / FilesLast Validated
flat-fileRecalls (FLAT_RCL.zip)Pending data
flat-fileComplaints (FLAT_CMPL.zip)Pending data
flat-fileInvestigations (FLAT_INV.zip)Pending data
flat-fileTSBs / Mfr Communications (chunked archives)Pending data
flat-fileManufacturer Communications bulkPending data
flat-fileRecall Completion RatesPending data
flat-fileData DictionariesPending data
flat-filevPIC Vehicle List (Lite)Pending data
flat-fileRecall Communications bulk PDFPending data
json-apiRecalls by VehiclePending data
json-apiComplaints by VehiclePending data
json-apiInvestigations APIPending data
json-apiSafety Ratings (NCAP)Pending data
json-apiManufacturer CommunicationsPending data
json-apiCSSI Station (Car Seat Inspection)Pending data
json-apivPIC Vehicle CatalogPending data
json-apiRecalls by Campaign NumberPending data
s3-pdf-listingRecall PDFs (S3 listing)Pending data
s3-pdf-listingTSB PDFs (S3 listing)Pending data
s3-pdf-listingInvestigation PDFs (S3 listing)Pending data
s3-pdf-listingMfr Comm PDFs (S3 listing)Pending data

Per-type contracts (NHV-1..7 validation gates) are written to docs/harvesting/nhtsa/data-types/ during Phases 2–5.