fix(wildlife): keep each failed source's retry timing after a partial failure

A partial failure reset the layer's cooldown, so a camera move retried at
once. Acartia's error was discarded, an absent X-Upstream-Status hid the
HTTP 503 (the proxy's daily-budget reply), and a grid schema error masked
the observation 503. Each failed source now keeps its status and
Retry-After, the HTTP status counts when no upstream status is sent, and the
layer waits for the longest 429/5xx cooldown among them. The shown error is
the one that explains the cooldown.

docs/CURRENT-STATE.md

@@ -124,7 +124,11 @@ with retry timing; logs exclude request URLs and upstream exception text.
 A partial failure still draws what arrived: a failed iNaturalist request, or
 one failed half of an antimeridian view, keeps a good Acartia reply and the
 other half, retains that area's last iNaturalist rows and reports DEGRADED.
-The update fails outright only when every requested source failed.
+The update fails outright only when every requested source failed. Each failed
+source keeps its HTTP status and `Retry-After` (the HTTP status counts when no
+`X-Upstream-Status` is sent), and the next request waits out the longest 429 or
+5xx cooldown among them. One update requests both sources, so the cooldown
+holds the whole layer; a fully successful refresh clears it.
 
 `/api/acartia/sightings` adds Salish Sea marine mammals within 45.5–50.5° N,
 126–121° W, cached ten minutes and stale for at most 24 h. A valid empty