feat(admin): add fleet statistics and admin read endpoints

Counters (always on when admin is enabled):
- last seen per user and group, written asynchronously after
  admission, throttled to one write per scope per minute per replica
- hourly provider and model attempts, failures and timeouts
- failovers per provider, rejections per hour, and unpriced (402)
  refusals
- cache hits, misses and cost saved
- request counts per target and caller

Opt-in counters under admin.stats (both default off):
- userModel: usage per user and model
- latency: latency histograms per provider and model

All counters ride the existing Redis batches. A test measures three
round trips per request with defaults and with every statistic on. A
cache hit takes two. A model request is no longer counted for failed
failover candidates.

New admin endpoints:
- usage/series and usage/totals
- performance, with percentiles from the histograms
- catalog, with the price billing actually uses and its source
- consumers, with key source, grants and model access
- config, with secrets redacted

The overview gains features and a pricing summary. Usage gains
lastSeen. usage/models gains offset and more detail fields. Reads are
chunked at 1600 keys per round trip and capped at 64000 keys per
request.

Claude-Session: https://claude.ai/code/session_01V5jYJNxpDFdJkyyExpADrv

events.go

@@ -452,6 +452,13 @@ func (g *Gateway) recordUnpricedRefusalEvent(scopes []limitScope, model, provide
 	}
 	applyScopeAttribution(&ev, scopes)
 	g.recordEvent(ev)
+	// r402 (always-on-with-admin, DECISIONS) — the one chokepoint both
+	// routes_unified.go's and routes_passthrough.go's identical 402 guard
+	// already share, so wiring the counter here covers both call sites
+	// without touching either route file again. Fire-and-forget
+	// (recordUnpriced402's own countAsync/l.spawn); no-op when admin
+	// stats are off.
+	g.limiter.recordUnpriced402(model)
 }
 
 // recordProxyEvent records an upstream event for a proxyUpstream result