Merged PR 41659: Add metrics with correct names for Resource Monitori…

…ng (#5341)

Add metrics with correct names for Resource Monitoring (#5341)

Fixes #5113
Previous art: #5309

Add new metrics with correct names. Old metrics will continue to be enabled by default.

### Existing metric setup

**Windows Snapshot provider class**
    `process.cpu.utilization`
    `dotnet.process.memory.virtual.utilization`

**Windows Container Snapshot provider class**
    `process.cpu.utilization`
    `dotnet.process.memory.virtual.utilization`

**Linix Utilization Provider class**
    `process.cpu.utilization`
    `dotnet.process.memory.virtual.utilization`

### New metric setup

**Windows Snapshot provider class**
    `process.cpu.utilization` - no changes
    `dotnet.process.memory.virtual.utilization` - no changes

**Windows Container Snapshot provider class**
    `process.cpu.utilization` - no changes
    `dotnet.process.memory.virtual.utilization` - calculates memory for the dotnet process only (instead of all processes)
    `container.cpu.limit.utilization` - new metric, same value as `process.cpu.utilization`
    `container.memory.limit.utilization` - new metric, calculates memory for all processes in the container

**Linux Utilization Provider class**
    `process.cpu.utilization` - fixed incorrect scale calculation, instead of `host CPUs / CPU limit / CPU request`, it is now `host CPUs / CPU request`
    `dotnet.process.memory.virtual.utilization` - no changes
    `container.cpu.limit.utilization` - new metric, value is relative to CPU resource limit (aka maximum CPU units)
    `container.memory.limit.utilization` - new metric, calculates memory for all processes in the container
    `container.cpu.request.utilization` - new metric, same value as `process.cpu.utilization`

----
#### AI description  (iteration 1)
#### PR Classification
New feature: Added metrics with correct names for resource monitoring.

#### PR Summary
This pull request introduces new metrics for resource monitoring with correct naming conventions and updates the related tests and implementation.
- `LinuxUtilizationProvider.cs`: Added new metrics for container CPU and memory utilization, and updated existing metrics.
- `AcceptanceTest.cs`: Added new tests for verifying the new metrics and updated existing tests for better coverage.
- `ResourceUtilizationInstruments.cs`: Defined new constants for the new metrics.
- Removed `WindowsCounters.cs` as it is no longer needed.

src/Libraries/Microsoft.Extensions.Diagnostics.ResourceMonitoring/Calculator.cs

@@ -25,7 +25,7 @@ public static ResourceUtilization CalculateUtilization(in Snapshot first, in Sna
         long runtimeTickDelta = second.TotalTimeSinceStart.Ticks - first.TotalTimeSinceStart.Ticks;
 
         // Compute the total number of ticks available on the machine during that interval
-        double totalSystemTicks = runtimeTickDelta * systemResources.GuaranteedCpuUnits;
+        double totalSystemTicks = runtimeTickDelta;
 
         // fudge to avoid divide by zero
         if (totalSystemTicks <= 0)

src/Libraries/Microsoft.Extensions.Diagnostics.ResourceMonitoring/Linux/LinuxUtilizationParserCgroupV1.cs

@@ -444,7 +444,7 @@ private static bool TryGetCpuUnitsFromCgroups(IFileSystem fileSystem, out float
     /// <summary>
     /// In cgroup v1 the CPU shares is used to determine the CPU allocation.
     /// in cgroup v2 the CPU weight is used to determine the CPU allocation.
-    /// To calculete CPU request in cgroup v2 we need to read the CPU weight and convert it to CPU shares.
+    /// To calculate CPU request in cgroup v2 we need to read the CPU weight and convert it to CPU shares.
     /// But for cgroup v1 we can read the CPU shares directly from the file.
     /// 1024 equals 1 CPU core.
     /// In cgroup v1 on some systems the location of the CPU shares file is different.

src/Libraries/Microsoft.Extensions.Diagnostics.ResourceMonitoring/Linux/LinuxUtilizationProvider.cs

@@ -15,12 +15,13 @@ internal sealed class LinuxUtilizationProvider : ISnapshotProvider
     private readonly object _cpuLocker = new();
     private readonly object _memoryLocker = new();
     private readonly ILinuxUtilizationParser _parser;
-    private readonly ulong _totalMemoryInBytes;
+    private readonly ulong _memoryLimit;
     private readonly TimeSpan _cpuRefreshInterval;
     private readonly TimeSpan _memoryRefreshInterval;
     private readonly TimeProvider _timeProvider;
-    private readonly double _scale;
-    private readonly double _scaleForTrackerApi;
+    private readonly double _scaleRelativeToCpuLimit;
+    private readonly double _scaleRelativeToCpuRequest;
+    private readonly double _scaleRelativeToCpuRequestForTrackerApi;
 
     private DateTimeOffset _refreshAfterCpu;
     private DateTimeOffset _refreshAfterMemory;
@@ -37,73 +38,73 @@ public LinuxUtilizationProvider(IOptions<ResourceMonitoringOptions> options, ILi
     {
         _parser = parser;
         _timeProvider = timeProvider ?? TimeProvider.System;
-        var now = _timeProvider.GetUtcNow();
+        DateTimeOffset now = _timeProvider.GetUtcNow();
         _cpuRefreshInterval = options.Value.CpuConsumptionRefreshInterval;
         _memoryRefreshInterval = options.Value.MemoryConsumptionRefreshInterval;
         _refreshAfterCpu = now;
         _refreshAfterMemory = now;
-        _totalMemoryInBytes = _parser.GetAvailableMemoryInBytes();
+        _memoryLimit = _parser.GetAvailableMemoryInBytes();
         _previousHostCpuTime = _parser.GetHostCpuUsageInNanoseconds();
         _previousCgroupCpuTime = _parser.GetCgroupCpuUsageInNanoseconds();
 
-        var hostMemory = _parser.GetHostAvailableMemory();
-        var hostCpus = _parser.GetHostCpuCount();
-        var availableCpus = _parser.GetCgroupLimitedCpus();
-        var cpuGuaranteedRequest = _parser.GetCgroupRequestCpu();
-        _scale = hostCpus / availableCpus;
-        _scaleForTrackerApi = hostCpus / availableCpus;
+        float hostCpus = _parser.GetHostCpuCount();
+        float cpuLimit = _parser.GetCgroupLimitedCpus();
+        float cpuRequest = _parser.GetCgroupRequestCpu();
+        _scaleRelativeToCpuLimit = hostCpus / cpuLimit;
+        _scaleRelativeToCpuRequest = hostCpus / cpuRequest;
+        _scaleRelativeToCpuRequestForTrackerApi = hostCpus; // the division by cpuRequest is performed later on in the ResourceUtilization class
 
 #pragma warning disable CA2000 // Dispose objects before losing scope
         // We don't dispose the meter because IMeterFactory handles that
         // An issue on analyzer side: https://github.com/dotnet/roslyn-analyzers/issues/6912
         // Related documentation: https://github.com/dotnet/docs/pull/37170
-        var meter = meterFactory.Create("Microsoft.Extensions.Diagnostics.ResourceMonitoring");
+        var meter = meterFactory.Create(nameof(Microsoft.Extensions.Diagnostics.ResourceMonitoring));
 #pragma warning restore CA2000 // Dispose objects before losing scope
 
-        _ = meter.CreateObservableGauge(name: ResourceUtilizationInstruments.CpuUtilization, observeValue: CpuUtilization, unit: "1");
-        _ = meter.CreateObservableGauge(name: ResourceUtilizationInstruments.MemoryUtilization, observeValue: MemoryUtilization, unit: "1");
+        _ = meter.CreateObservableGauge(name: ResourceUtilizationInstruments.ContainerCpuLimitUtilization, observeValue: () => CpuUtilization() * _scaleRelativeToCpuLimit, unit: "1");
+        _ = meter.CreateObservableGauge(name: ResourceUtilizationInstruments.ContainerMemoryLimitUtilization, observeValue: MemoryUtilization, unit: "1");
+        _ = meter.CreateObservableGauge(name: ResourceUtilizationInstruments.ContainerCpuRequestUtilization, observeValue: () => CpuUtilization() * _scaleRelativeToCpuRequest, unit: "1");
 
-        // cpuGuaranteedRequest is a CPU request for pod, for host its 1 core
-        // available CPUs is a CPU limit for a pod or for a host.
-        // _totalMemoryInBytes - Resource Memory Limit (in k8s terms)
-        // _totalMemoryInBytes - To keep the contract, this parameter will get the Host available memory
-        Resources = new SystemResources(cpuGuaranteedRequest, availableCpus, _totalMemoryInBytes, _totalMemoryInBytes);
+        _ = meter.CreateObservableGauge(name: ResourceUtilizationInstruments.ProcessCpuUtilization, observeValue: () => CpuUtilization() * _scaleRelativeToCpuRequest, unit: "1");
+        _ = meter.CreateObservableGauge(name: ResourceUtilizationInstruments.ProcessMemoryUtilization, observeValue: MemoryUtilization, unit: "1");
+
+        // cpuRequest is a CPU request (aka guaranteed number of CPU units) for pod, for host its 1 core
+        // cpuLimit is a CPU limit (aka max CPU units available) for a pod or for a host.
+        // _memoryLimit - Resource Memory Limit (in k8s terms)
+        // _memoryLimit - To keep the contract, this parameter will get the Host available memory
+        Resources = new SystemResources(cpuRequest, cpuLimit, _memoryLimit, _memoryLimit);
     }
 
     public double CpuUtilization()
     {
-        var now = _timeProvider.GetUtcNow();
-        bool needUpdate = false;
+        DateTimeOffset now = _timeProvider.GetUtcNow();
 
         lock (_cpuLocker)
         {
-            if (now >= _refreshAfterCpu)
+            if (now < _refreshAfterCpu)
             {
-                needUpdate = true;
+                return _cpuPercentage;
             }
         }
 
-        if (needUpdate)
-        {
-            var hostCpuTime = _parser.GetHostCpuUsageInNanoseconds();
-            var cgroupCpuTime = _parser.GetCgroupCpuUsageInNanoseconds();
+        long hostCpuTime = _parser.GetHostCpuUsageInNanoseconds();
+        long cgroupCpuTime = _parser.GetCgroupCpuUsageInNanoseconds();
 
-            lock (_cpuLocker)
+        lock (_cpuLocker)
+        {
+            if (now >= _refreshAfterCpu)
             {
-                if (now >= _refreshAfterCpu)
+                double deltaHost = hostCpuTime - _previousHostCpuTime;
+                double deltaCgroup = cgroupCpuTime - _previousCgroupCpuTime;
+
+                if (deltaHost > 0 && deltaCgroup > 0)
                 {
-                    var deltaHost = hostCpuTime - _previousHostCpuTime;
-                    var deltaCgroup = cgroupCpuTime - _previousCgroupCpuTime;
-
-                    if (deltaHost > 0 && deltaCgroup > 0)
-                    {
-                        var percentage = Math.Min(One, deltaCgroup / deltaHost * _scale);
-
-                        _cpuPercentage = percentage;
-                        _refreshAfterCpu = now.Add(_cpuRefreshInterval);
-                        _previousCgroupCpuTime = cgroupCpuTime;
-                        _previousHostCpuTime = hostCpuTime;
-                    }
+                    double percentage = Math.Min(One, deltaCgroup / deltaHost);
+
+                    _cpuPercentage = percentage;
+                    _refreshAfterCpu = now.Add(_cpuRefreshInterval);
+                    _previousCgroupCpuTime = cgroupCpuTime;
+                    _previousHostCpuTime = hostCpuTime;
                 }
             }
         }
@@ -113,30 +114,26 @@ public double CpuUtilization()
 
     public double MemoryUtilization()
     {
-        var now = _timeProvider.GetUtcNow();
-        bool needUpdate = false;
+        DateTimeOffset now = _timeProvider.GetUtcNow();
 
         lock (_memoryLocker)
         {
-            if (now >= _refreshAfterMemory)
+            if (now < _refreshAfterMemory)
             {
-                needUpdate = true;
+                return _memoryPercentage;
             }
         }
 
-        if (needUpdate)
-        {
-            var memoryUsed = _parser.GetMemoryUsageInBytes();
+        ulong memoryUsed = _parser.GetMemoryUsageInBytes();
 
-            lock (_memoryLocker)
+        lock (_memoryLocker)
+        {
+            if (now >= _refreshAfterMemory)
             {
-                if (now >= _refreshAfterMemory)
-                {
-                    var memoryPercentage = Math.Min(One, (double)memoryUsed / _totalMemoryInBytes);
+                double memoryPercentage = Math.Min(One, (double)memoryUsed / _memoryLimit);
 
-                    _memoryPercentage = memoryPercentage;
-                    _refreshAfterMemory = now.Add(_memoryRefreshInterval);
-                }
+                _memoryPercentage = memoryPercentage;
+                _refreshAfterMemory = now.Add(_memoryRefreshInterval);
             }
         }
 
@@ -150,14 +147,14 @@ public double MemoryUtilization()
     /// </remarks>
     public Snapshot GetSnapshot()
     {
-        var hostTime = _parser.GetHostCpuUsageInNanoseconds();
-        var cgroupTime = _parser.GetCgroupCpuUsageInNanoseconds();
-        var memoryUsed = _parser.GetMemoryUsageInBytes();
+        long hostTime = _parser.GetHostCpuUsageInNanoseconds();
+        long cgroupTime = _parser.GetCgroupCpuUsageInNanoseconds();
+        ulong memoryUsed = _parser.GetMemoryUsageInBytes();
 
         return new Snapshot(
             totalTimeSinceStart: TimeSpan.FromTicks(hostTime / Hundred),
             kernelTimeSinceStart: TimeSpan.Zero,
-            userTimeSinceStart: TimeSpan.FromTicks((long)(cgroupTime / Hundred * _scaleForTrackerApi)),
+            userTimeSinceStart: TimeSpan.FromTicks((long)(cgroupTime / Hundred * _scaleRelativeToCpuRequestForTrackerApi)),
             memoryUsageInBytes: memoryUsed);
     }
 }

src/Libraries/Microsoft.Extensions.Diagnostics.ResourceMonitoring/ResourceMonitoringServiceCollectionExtensions.cs

@@ -88,7 +88,7 @@ private static ResourceMonitorBuilder AddWindowsProvider(this ResourceMonitorBui
         builder.PickWindowsSnapshotProvider();
 
         _ = builder.Services
-            .AddActivatedSingleton<WindowsCounters>();
+            .AddActivatedSingleton<WindowsNetworkMetrics>();
 
         _ = builder.Services
             .AddActivatedSingleton<TcpTableInfo>();

src/Libraries/Microsoft.Extensions.Diagnostics.ResourceMonitoring/ResourceUtilization.cs

@@ -51,10 +51,16 @@ public readonly struct ResourceUtilization
     /// <param name="systemResources">CPU and memory limits.</param>
     public ResourceUtilization(double cpuUsedPercentage, ulong memoryUsedInBytes, SystemResources systemResources)
     {
-        CpuUsedPercentage = Throw.IfLessThan(cpuUsedPercentage, 0.0);
+        double guaranteedCpuUnits = systemResources.GuaranteedCpuUnits;
+        if (guaranteedCpuUnits <= 0)
+        {
+            guaranteedCpuUnits = 1;
+        }
+
+        CpuUsedPercentage = Throw.IfLessThan(cpuUsedPercentage / guaranteedCpuUnits, 0.0);
         MemoryUsedInBytes = Throw.IfLessThan(memoryUsedInBytes, 0);
         SystemResources = systemResources;
-        MemoryUsedPercentage = Math.Min(Hundred, (double)MemoryUsedInBytes / SystemResources.GuaranteedMemoryInBytes * Hundred);
+        MemoryUsedPercentage = Math.Min(Hundred, (double)MemoryUsedInBytes / systemResources.GuaranteedMemoryInBytes * Hundred);
     }
 
     /// <summary>

src/Libraries/Microsoft.Extensions.Diagnostics.ResourceMonitoring/ResourceUtilizationInstruments.cs

@@ -1,6 +1,8 @@
 // Licensed to the .NET Foundation under one or more agreements.
 // The .NET Foundation licenses this file to you under the MIT license.
 
+using System;
+
 namespace Microsoft.Extensions.Diagnostics.ResourceMonitoring;
 
 /// <summary>
@@ -13,18 +15,42 @@ namespace Microsoft.Extensions.Diagnostics.ResourceMonitoring;
 internal static class ResourceUtilizationInstruments
 {
     /// <summary>
-    /// Gets the CPU consumption of the running application in range <c>[0, 1]</c>.
+    /// The name of an instrument to retrieve CPU limit consumption of all processes running inside a container or control group in range <c>[0, 1]</c>.
+    /// </summary>
+    /// <remarks>
+    /// The type of an instrument is <see cref="System.Diagnostics.Metrics.ObservableGauge{T}"/>.
+    /// </remarks>
+    public const string ContainerCpuLimitUtilization = "container.cpu.limit.utilization";
+
+    /// <summary>
+    /// The name of an instrument to retrieve CPU request consumption of all processes running inside a container or control group in range <c>[0, 1]</c>.
+    /// </summary>
+    /// <remarks>
+    /// The type of an instrument is <see cref="System.Diagnostics.Metrics.ObservableGauge{T}"/>.
+    /// </remarks>
+    public const string ContainerCpuRequestUtilization = "container.cpu.request.utilization";
+
+    /// <summary>
+    /// The name of an instrument to retrieve memory limit consumption of all processes running inside a container or control group in range <c>[0, 1]</c>.
+    /// </summary>
+    /// <remarks>
+    /// The type of an instrument is <see cref="System.Diagnostics.Metrics.ObservableGauge{T}"/>.
+    /// </remarks>
+    public const string ContainerMemoryLimitUtilization = "container.memory.limit.utilization";
+
+    /// <summary>
+    /// The name of an instrument to retrieve CPU consumption share of the running process in range <c>[0, 1]</c>.
     /// </summary>
     /// <remarks>
     /// The type of an instrument is <see cref="System.Diagnostics.Metrics.ObservableGauge{T}"/>.
     /// </remarks>
-    public const string CpuUtilization = "process.cpu.utilization";
+    public const string ProcessCpuUtilization = "process.cpu.utilization";
 
     /// <summary>
-    /// Gets the memory consumption of the running application in range <c>[0, 1]</c>.
+    /// The name of an instrument to retrieve memory consumption share of the running process in range <c>[0, 1]</c>.
     /// </summary>
     /// <remarks>
     /// The type of an instrument is <see cref="System.Diagnostics.Metrics.ObservableGauge{T}"/>.
     /// </remarks>
-    public const string MemoryUtilization = "dotnet.process.memory.virtual.utilization";
+    public const string ProcessMemoryUtilization = "dotnet.process.memory.virtual.utilization";
 }

src/Libraries/Microsoft.Extensions.Diagnostics.ResourceMonitoring/Windows/Interop/IProcessInfo.cs

@@ -9,8 +9,14 @@ namespace Microsoft.Extensions.Diagnostics.ResourceMonitoring.Windows.Interop;
 internal interface IProcessInfo
 {
     /// <summary>
-    /// Retrieve the memory usage of a system.
+    /// Retrieves the amount of memory, in bytes, used by the current process.
     /// </summary>
-    /// <returns>Memory usage amount in bytes.</returns>
+    /// <returns>The number of bytes allocated by the current process.</returns>
+    ulong GetCurrentProcessMemoryUsage();
+
+    /// <summary>
+    /// Retrieves the amount of memory, in bytes, used by the system.
+    /// </summary>
+    /// <returns>The number of bytes allocated by the system.</returns>
     ulong GetMemoryUsage();
 }

src/Libraries/Microsoft.Extensions.Diagnostics.ResourceMonitoring/Windows/Interop/ProcessInfo.cs

@@ -38,4 +38,10 @@ public ulong GetMemoryUsage()
 
         return memoryUsage;
     }
+
+    public ulong GetCurrentProcessMemoryUsage()
+    {
+        using Process process = Process.GetCurrentProcess();
+        return (ulong)process.WorkingSet64;
+    }
 }