perf(metrics): tiered rollup history (7d at ~1.5MB) and cleaner ranges

Replace the flat 48h@2s ring buffer with a 3-tier rollup ladder (2s/1h, 1m/48h, 10m/7d). A sample feeds every tier and rolls up into progressively coarser averages, so per-metric footprint drops from ~21MB to ~1.5MB (measured, 16 system metrics) while extending the range from 48h to 7 days. aggregate() picks the finest tier covering the requested span; a pre-tier flat gob is migrated by replaying its samples through the rollup.

Tidy the dashboard ranges to a professional ladder: 2m, 1h, 3h, 6h, 12h, 24h, 2d, 7d (drop the irregular 2h/5h, the redundant 30m, and the excessive 30d). The allow-list keeps bucket 30 because the node history panel uses it.

Add an initial FreeOSMemory about 60s after boot to reclaim the startup and metric-restore peak instead of waiting for the periodic release. Cover the rollup, tier selection, round-trip, and footprint with tests.

frontend/src/pages/index/XrayMetricsModal.tsx

@@ -286,11 +286,10 @@ export default function XrayMetricsModal({ open, onClose }: XrayMetricsModalProp
             onChange={setBucket}
             options={[
               { value: 2, label: '2m' },
-              { value: 30, label: '30m' },
               { value: 60, label: '1h' },
-              { value: 120, label: '2h' },
               { value: 180, label: '3h' },
-              { value: 300, label: '5h' },
+              { value: 360, label: '6h' },
+              { value: 720, label: '12h' },
             ]}
           />
         </div>