fix(unraid): report correct CPU utilization on hyper-threaded systems

The Unraid integration summed the per-thread CPU percentages and divided
the total by the number of physical cores. On hyper-threaded systems there
are more logical CPUs (threads) than physical cores, so the result was
inflated — roughly doubled — and could even exceed 100% (e.g. a real 43%
load reported as ~93%).

Use the overall CPU utilization the Unraid API already reports
(`metrics.cpu.percentTotal`, 0-100), which is the value Unraid itself
displays, instead of recomputing it from the per-thread percentages.

Closes #6082

Author: junkerderprovinz

packages/integrations/src/unraid/unraid-integration.ts

@@ -37,9 +37,6 @@ export class UnraidIntegration extends Integration implements ISystemHealthMonit
   public async getSystemInfoAsync(): Promise<SystemHealthMonitoring> {
     const systemInfo = await this.getSystemInformationAsync();
 
-    const cpuUtilization = systemInfo.metrics.cpu.cpus.reduce((acc, val) => acc + val.percentTotal, 0);
-    const cpuCount = systemInfo.info.cpu.cores;
-
     // We use "info" object instead of the stats since this is the exact amount the kernel sees, which is what Unraid displays.
     const totalMemory = systemInfo.info.memory.layout.reduce((acc, layout) => layout.size + acc, 0);
     const usedMemory = totalMemory * (systemInfo.metrics.memory.percentTotal / 100);
@@ -48,7 +45,12 @@ export class UnraidIntegration extends Integration implements ISystemHealthMonit
     return {
       version: systemInfo.info.os.release,
       cpuModelName: systemInfo.info.cpu.brand,
-      cpuUtilization: cpuUtilization / cpuCount,
+      // Use the overall CPU utilization the Unraid API already reports (0-100),
+      // which is the value Unraid itself displays. Previously the per-thread
+      // percentages were summed and divided by the number of physical cores,
+      // which double-counted on hyper-threaded systems (e.g. 24 threads over
+      // 12 cores) and could even exceed 100%.
+      cpuUtilization: systemInfo.metrics.cpu.percentTotal,
       memUsedInBytes: usedMemory,
       memAvailableInBytes: totalMemory - usedMemory,
       uptime: dayjs().diff(uptime, "seconds"),