fix(geocoding): bump PROVIDER_TIMEOUT_MS default 5s → 8s

First-probe DNS+TLS handshake against Nominatim can take >5s on a
cold start (verified locally: 642ms warm, sometimes 5-8s cold). The
old 5s default false-marked Nominatim unhealthy and the 30s health-
cache then locked us into a fallback-of-fallback gap. 8s gives
enough headroom for cold-start while still cutting off actually-
stuck connections.

Photon and Pelias don't hit this — Photon's CDN is consistently
sub-second and Pelias is on localhost / LAN. Only the public
Nominatim path warranted the bump, but the timeout is per-provider
shared so we adjust it globally.

Existing PROVIDER_TIMEOUT_MS env override still wins.

services/mana-geocoding/CLAUDE.md

@@ -154,7 +154,7 @@ PORT=3018
 
 # --- Provider chain (tried in order) ----------------------------------
 GEOCODING_PROVIDERS=pelias,photon,nominatim
-PROVIDER_TIMEOUT_MS=5000              # per-provider request timeout
+PROVIDER_TIMEOUT_MS=8000              # per-provider request timeout (cold-start safe)
 PROVIDER_HEALTH_CACHE_MS=30000        # health-cache TTL — skip dead providers
 
 # --- Pelias (primary) -------------------------------------------------

services/mana-geocoding/src/config.ts

@@ -72,7 +72,11 @@ export function loadConfig(): Config {
 				'nominatim',
 			]),
 			healthCacheMs: parseInt(process.env.PROVIDER_HEALTH_CACHE_MS || '30000', 10),
-			timeoutMs: parseInt(process.env.PROVIDER_TIMEOUT_MS || '5000', 10),
+			// 8 s default. Nominatim's cold-start DNS+TLS handshake can push the
+			// first health probe past the older 5 s default, false-marking the
+			// provider unhealthy for the next 30 s. 8 s survives a slow first
+			// probe but still cuts off actually-stuck connections.
+			timeoutMs: parseInt(process.env.PROVIDER_TIMEOUT_MS || '8000', 10),
 		},
 	};
 }