STRATA · Pipeline

Fetcher health

Live state of the 132 sources on automated fetch - a subset of the full source catalogue, the rest being curated, vendored or refreshed by hand. Which caches are present and which parse cleanly into records, generated by the strata probe.

Data pipeline · source fleet

Fetcher health

Sources

132

Healthy

131

Broken

Un-fetched

Records parsed

2,531,030

131 + 0 + 1 = 132 sources. 35 facility · 97 commodity · 101 parse-verified · 24 large · 3 staging-only · 0 empty-cache. Generated 2026-09-15 16:15 UTC.

Facility domain · 35 sources

Facility & entity backbone

Source	Vintage	Status	Records	Size
brightquery_global	2026-06	Un-fetched	-	-
asx_substantial probe timed out after 120s (live re-validation?)	2026-08-15	Present (live)	-	25.6 MB
anm_sigmine	2026	Present (large)	-

Commodity domain · 97 sources

Commodity statistics & deposits

Source	Vintage	Status	Records	Size
sec_edgar_ex96 probe timed out after 120s (live re-validation?)	2026-06-16	Present (live)	-	163 KB
usmin_critical_deposits probe timed out after 120s (live re-validation?)	2024	Present (live)	-	583 KB
anm_cfem	2026	Present (large)

Provenance

How this is measured

per-source cache state + a fetch()->parse() smoke-test counting emitted records on every source with a cached artifact <= 150 MB; larger artifacts are confirmed present but not re-parsed inline. Staging-only lanes emit no record by design, so they are reported as such rather than smoke-tested. A source is facility-domain iff every record kind it emits is in the facility canonical store; otherwise it is commodity-domain. This view is generated by strata sources health and regenerated with pnpm health:refresh - never hand-edited.