feat: add server stats, GPU stats, container CPU/memory display

- Add /api/servers endpoint querying Prometheus for CPU, RAM, disk, uptime, load
- Add /api/gpu endpoint for NVIDIA Jetson GPU utilization, temp, power
- Add ServerStatsWidget with animated bars for Atlas, Wile, RoadRunner
- Add GPUStatsWidget with GPU util, memory, temp color-coding, power draw
- Update ContainerGroup to show CPU bar and memory for running containers
- Fix docker-compose.yml: traefik network external: true
- Fix getTraefikUrl to scan all router labels (not just 'https')

src/app/api/gpu/route.ts

@@ -0,0 +1,87 @@
+import { NextResponse } from "next/server";
+
+const PROMETHEUS_URL = "http://prometheus:9090";
+
+const INSTANCE_MAP: Record<string, string> = {
+  "192.168.1.50": "Wile",
+  "192.168.1.51": "RoadRunner",
+};
+
+async function queryPrometheus(query: string): Promise<any[]> {
+  try {
+    const url = `${PROMETHEUS_URL}/api/v1/query?query=${encodeURIComponent(query)}`;
+    const res = await fetch(url, { cache: "no-store" });
+    if (!res.ok) return [];
+    const json = await res.json();
+    if (json.status === "success" && json.data?.result) {
+      return json.data.result;
+    }
+    return [];
+  } catch {
+    return [];
+  }
+}
+
+function extractByInstance(results: any[]): Record<string, number> {
+  const map: Record<string, number> = {};
+  for (const r of results) {
+    const instance: string = r.metric?.instance || "";
+    const ip = instance.replace(/:\d+$/, "");
+    const val = parseFloat(r.value?.[1] || "0");
+    if (!isNaN(val)) {
+      map[ip] = val;
+    }
+  }
+  return map;
+}
+
+async function queryWithFallback(
+  primaryMetric: string,
+  ...fallbacks: string[]
+): Promise<Record<string, number>> {
+  const primary = await queryPrometheus(primaryMetric);
+  if (primary.length > 0) return extractByInstance(primary);
+
+  for (const fb of fallbacks) {
+    const res = await queryPrometheus(fb);
+    if (res.length > 0) return extractByInstance(res);
+  }
+  return {};
+}
+
+export async function GET() {
+  try {
+    const [gpuUtilMap, memUtilMap, tempMap, powerMap] = await Promise.all([
+      queryWithFallback(
+        "DCGM_FI_DEV_GPU_UTIL",
+        "nvidia_gpu_utilization_gpu",
+        "gpu_utilization_percentage"
+      ),
+      queryWithFallback(
+        "DCGM_FI_DEV_MEM_COPY_UTIL",
+        "nvidia_gpu_memory_used_bytes / nvidia_gpu_memory_total_bytes * 100"
+      ),
+      queryWithFallback(
+        "DCGM_FI_DEV_GPU_TEMP",
+        "nvidia_gpu_temperature_gpu"
+      ),
+      queryWithFallback(
+        "DCGM_FI_DEV_POWER_USAGE",
+        "nvidia_gpu_power_draw_watts"
+      ),
+    ]);
+
+    const gpus = Object.entries(INSTANCE_MAP).map(([ip, name]) => ({
+      name,
+      gpu_util: parseFloat((gpuUtilMap[ip] || 0).toFixed(1)),
+      mem_util: parseFloat((memUtilMap[ip] || 0).toFixed(1)),
+      temp: parseFloat((tempMap[ip] || 0).toFixed(0)),
+      power_watts: parseFloat((powerMap[ip] || 0).toFixed(1)),
+    }));
+
+    return NextResponse.json(gpus);
+  } catch (error) {
+    console.error("GPU API error:", error);
+    return NextResponse.json([]);
+  }
+}