feat: realtime Live dashboard in Data Explorer + fix panel layout

- New "Live" tab: auto-polls /api/federated/live every 2.5s with
  animated counters, ingestion throughput sparkline, per-source
  write-rate bars, a region scorecard matrix (heat-shaded) and live
  business breakdown charts (region/channel/status/customers/
  telemetry/supply)
- Backend /api/federated/live: instant source estimates (Postgres),
  monotonic max(event_id) for MySQL and Mongo estimated count for
  immediate movement, Cassandra from cached matrix; business aggs
  cached over the small Hadoop lake tables (short TTL)
- Fix embedded Trino panels being squeezed with internal scrollbars
  by making panels/grids shrink-0 so the page scrolls instead

api/trino_federated.py

@@ -235,6 +235,116 @@ async def get_catalogs():
     return {"ok": True, "catalogs": out, "source_totals": totals, "count": len(out)}
 
 
+# ──────────────────────────────────────────────────────────────────────────────
+# Realtime business dashboard — fast: instant source estimates + short-TTL
+# cached aggregations over the small materialized Hadoop lake tables.
+# ──────────────────────────────────────────────────────────────────────────────
+_live_aggs: dict[str, Any] = {"ts": 0.0, "data": None}
+_LIVE_AGG_TTL = 6.0
+_avg_order_cache: dict[str, Any] = {"ts": 0.0, "val": 0.0}
+
+
+def _avg_order_value() -> float:
+    now = time.time()
+    if now - _avg_order_cache["ts"] < 300 and _avg_order_cache["val"]:
+        return _avg_order_cache["val"]
+    res = _trino("SELECT avg(amount) FROM iceberg.hadoop.orders_ext", 1)
+    val = 0.0
+    if res.get("ok"):
+        try:
+            val = float((res.get("rows") or [[0]])[0][0] or 0)
+        except Exception:
+            val = 0.0
+    _avg_order_cache["val"] = val
+    _avg_order_cache["ts"] = now
+    return val
+
+
+def _region_matrix_from_lake() -> list[dict]:
+    o = _terms("SELECT region k, count(*) c, sum(amount) rev FROM iceberg.hadoop.orders_ext GROUP BY region", "k", "c", "rev")
+    e = _terms("SELECT region k, count(*) c FROM iceberg.hadoop.employees_ext GROUP BY region", "k", "c")
+    sup = _terms("SELECT region k, count(*) c FROM iceberg.hadoop.supply_events_ext GROUP BY region", "k", "c")
+    regions: dict[str, dict] = {}
+    for x in o:
+        regions.setdefault(x["key"], {})["orders"] = x["count"]
+        regions[x["key"]]["revenue"] = x.get("value", 0)
+    for x in e:
+        regions.setdefault(x["key"], {})["hr_events"] = x["count"]
+    for x in sup:
+        regions.setdefault(x["key"], {})["supply_events"] = x["count"]
+    rows = [{"region": k, "orders": v.get("orders", 0), "revenue": v.get("revenue", 0),
+             "hr_events": v.get("hr_events", 0), "supply_events": v.get("supply_events", 0)}
+            for k, v in regions.items() if k]
+    rows.sort(key=lambda r: r.get("revenue") or 0, reverse=True)
+    return rows
+
+
+def _live_business_aggs() -> dict[str, Any]:
+    now = time.time()
+    if _live_aggs["data"] is not None and now - _live_aggs["ts"] < _LIVE_AGG_TTL:
+        return _live_aggs["data"]
+    data = {
+        "orders_by_region": _terms("SELECT region, count(*) c, sum(amount) rev FROM iceberg.hadoop.orders_ext GROUP BY region ORDER BY rev DESC", "region", "c", "rev"),
+        "orders_by_status": _terms("SELECT order_status k, count(*) c FROM iceberg.hadoop.orders_ext GROUP BY order_status ORDER BY c DESC", "k", "c"),
+        "orders_by_channel": _terms("SELECT sales_channel k, count(*) c, sum(amount) rev FROM iceberg.hadoop.orders_ext GROUP BY sales_channel ORDER BY rev DESC", "k", "c", "rev"),
+        "top_customers": _terms("SELECT customer_name k, count(*) c, sum(amount) rev FROM iceberg.hadoop.orders_ext GROUP BY customer_name ORDER BY rev DESC LIMIT 8", "k", "c", "rev"),
+        "telemetry_by_metric": _terms("SELECT metric_type k, count(*) c, avg(metric_value) v FROM iceberg.hadoop.telemetry_ext GROUP BY metric_type ORDER BY c DESC", "k", "c", "v"),
+        "supply_by_type": _terms("SELECT type k, count(*) c FROM iceberg.hadoop.supply_events_ext GROUP BY type ORDER BY c DESC", "k", "c"),
+        "region_matrix": _region_matrix_from_lake(),
+    }
+    _live_aggs["data"] = data
+    _live_aggs["ts"] = now
+    return data
+
+
+@router.get("/live")
+async def get_live():
+    import sql_console as s
+    orders = s._table_row_count("postgres", "public.sales_orders") or 0
+    # MySQL event_id is monotonic, so max(event_id) tracks inserts in real time
+    # (the planner estimate only refreshes after ANALYZE).
+    hr_res = _trino("SELECT max(event_id) FROM mysql_hr.hr.employee_events", 1)
+    hr = 0
+    if hr_res.get("ok"):
+        try:
+            hr = int((hr_res.get("rows") or [[0]])[0][0] or 0)
+        except Exception:
+            hr = 0
+    if not hr:
+        hr = s._table_row_count("mysql", "hr.employee_events") or 0
+    supply = s._table_row_count("mongodb", "supplychain.events") or 0
+    # Cassandra has no cheap estimate — reuse the exact count from the cached
+    # federated matrix query when available.
+    telemetry = 0
+    if _marquee.get("data") is None:
+        _load_marquee()
+    mq = _marquee.get("data") or {}
+    for r in ((mq.get("matrix") or {}).get("rows") or []):
+        if r.get("catalog") == "cassandra_telemetry":
+            try:
+                telemetry = int(r.get("records") or 0)
+            except Exception:
+                telemetry = 0
+    avg_order = _avg_order_value()
+    sources = [
+        {"key": "orders", "label": "Orders", "engine": "PostgreSQL", "catalog": "postgres_sales", "rows": orders, "color": "#fbbf24"},
+        {"key": "hr_events", "label": "HR events", "engine": "MySQL", "catalog": "mysql_hr", "rows": hr, "color": "#60a5fa"},
+        {"key": "supply_events", "label": "Supply events", "engine": "MongoDB", "catalog": "mongodb_supplychain", "rows": supply, "color": "#a78bfa"},
+        {"key": "telemetry", "label": "Telemetry", "engine": "Cassandra", "catalog": "cassandra_telemetry", "rows": telemetry, "color": "#22d3ee"},
+    ]
+    return {
+        "ok": True,
+        "ts": datetime.now(timezone.utc).isoformat(),
+        "sources": sources,
+        "totals": {
+            "records": orders + hr + supply + telemetry,
+            "revenue_est": round(orders * avg_order, 2),
+            "avg_order": round(avg_order, 2),
+        },
+        "business": _live_business_aggs(),
+    }
+
+
 # ──────────────────────────────────────────────────────────────────────────────
 # Hadoop lake analytics (live over the small materialized external tables)
 # ──────────────────────────────────────────────────────────────────────────────