[receiver/vcenter] add vcenter vm cpu usage metrics (open-telemetry#2…

…0894)

* add vcenter vm cpu usage metrics

.chloggen/vm_cpu.yaml

@@ -0,0 +1,16 @@
+# One of 'breaking', 'deprecation', 'new_component', 'enhancement', 'bug_fix'
+change_type: enhancement
+
+# The name of the component, or a single word describing the area of concern, (e.g. filelogreceiver)
+component: vcenterreceiver
+
+# A brief description of the change. Surround your text with quotes ("") if it needs to start with a backtick (`).
+note: Adds VM CPU usage and utilization metrics
+
+# One or more tracking issues related to the change
+issues: [20895]
+
+# (Optional) One or more lines of additional information to render under the primary note.
+# These lines will be padded with 2 spaces and then inserted directly into the document.
+# Use pipe (|) for multiline entries.
+subtext:

receiver/vcenterreceiver/documentation.md

@@ -262,6 +262,22 @@ The usage of the memory by the resource pool.
 | ---- | ----------- | ---------- | ----------------------- | --------- |
 | MiBy | Sum | Int | Cumulative | false |
 
+### vcenter.vm.cpu.usage
+
+The amount of CPU used by the VM.
+
+| Unit | Metric Type | Value Type | Aggregation Temporality | Monotonic |
+| ---- | ----------- | ---------- | ----------------------- | --------- |
+| MHz | Sum | Int | Cumulative | false |
+
+### vcenter.vm.cpu.utilization
+
+The CPU utilization of the VM.
+
+| Unit | Metric Type | Value Type |
+| ---- | ----------- | ---------- |
+| % | Gauge | Double |
+
 ### vcenter.vm.disk.latency.avg
 
 The latency of operations to the virtual machine's disk.

receiver/vcenterreceiver/internal/metadata/testdata/config.yaml

@@ -49,6 +49,10 @@ all_set:
  enabled: true
  vcenter.resource_pool.memory.usage:
  enabled: true
+ vcenter.vm.cpu.usage:
+ enabled: true
+ vcenter.vm.cpu.utilization:
+ enabled: true
  vcenter.vm.disk.latency.avg:
  enabled: true
  vcenter.vm.disk.latency.max:
@@ -136,6 +140,10 @@ none_set:
  enabled: false
  vcenter.resource_pool.memory.usage:
  enabled: false
+ vcenter.vm.cpu.usage:
+ enabled: false
+ vcenter.vm.cpu.utilization:
+ enabled: false
  vcenter.vm.disk.latency.avg:
  enabled: false
  vcenter.vm.disk.latency.max:

receiver/vcenterreceiver/internal/mockserver/README.md

@@ -35,13 +35,12 @@ And then running the receiver against the proxy.
 ```yaml
 receivers:
  vcenter:
- metrics:
- endpoint: https://localhost:56626
- username: "otelu"
- password: "otelp"
- tls:
- insecure: false
- insecure_skip_verify: true
+ endpoint: https://localhost:56626
+ username: "otelu"
+ password: "otelp"
+ tls:
+ insecure: false
+ insecure_skip_verify: true
 
 service:
  pipelines:

receiver/vcenterreceiver/internal/mockserver/client_mock.go

@@ -184,6 +184,9 @@ func routeRetreiveProperties(t *testing.T, body map[string]interface{}) ([]byte,
  if ps == "name" {
  return loadResponse("host-names.xml")
  }
+ if ps == "summary.hardware" {
+ return loadResponse("host-properties.xml")
+ }
 
  }
 

receiver/vcenterreceiver/internal/mockserver/responses/vm-properties.xml

@@ -1268,7 +1268,7 @@
  <timestamp>2022-05-17T14:28:26.159606Z</timestamp>
  </storage>
  <quickStats>
- <overallCpuUsage>0</overallCpuUsage>
+ <overallCpuUsage>12</overallCpuUsage>
  <overallCpuDemand>0</overallCpuDemand>
  <overallCpuReadiness>0</overallCpuReadiness>
  <guestMemoryUsage>163</guestMemoryUsage>

receiver/vcenterreceiver/metadata.yaml

@@ -382,3 +382,19 @@ metrics:
  aggregation: cumulative
  attributes: []
  extended_documentation: As measured over the most recent 20s interval.
+ vcenter.vm.cpu.utilization:
+ enabled: true
+ description: The CPU utilization of the VM.
+ unit: "%"
+ gauge:
+ value_type: double
+ attributes: []
+ vcenter.vm.cpu.usage:
+ enabled: true
+ description: The amount of CPU used by the VM.
+ unit: "MHz"
+ sum:
+ monotonic: false
+ value_type: int
+ aggregation: cumulative
+ attributes: []

receiver/vcenterreceiver/metrics.go

@@ -51,6 +51,7 @@ func (v *vcenterMetricScraper) recordHostSystemMemoryUsage(
 func (v *vcenterMetricScraper) recordVMUsages(
  now pcommon.Timestamp,
  vm mo.VirtualMachine,
+ hs mo.HostSystem,
 ) {
  memUsage := vm.Summary.QuickStats.GuestMemoryUsage
  balloonedMem := vm.Summary.QuickStats.BalloonedMemory
@@ -71,6 +72,33 @@ func (v *vcenterMetricScraper) recordVMUsages(
  diskUtilization := float64(diskUsed) / float64(diskFree+diskUsed) * 100
  v.mb.RecordVcenterVMDiskUtilizationDataPoint(now, diskUtilization)
  }
+
+ s := vm.Summary
+ z := s.QuickStats
+
+ cpuUsage := z.OverallCpuUsage
+ ncpu := vm.Config.Hardware.NumCPU
+
+ // if no cpu usage, we probably shouldn't return a value. Most likely the VM is unavailable
+ // or is unreachable.
+ if cpuUsage == 0 {
+ return
+ }
+
+ // https://communities.vmware.com/t5/VMware-code-Documents/Resource-Management/ta-p/2783456
+ // VirtualMachine.runtime.maxCpuUsage is a property of the virtual machine, indicating the limit value.
+ // This value is always equal to the limit value set for that virtual machine.
+ // If no limit, it has full host mhz * vm.Config.Hardware.NumCPU.
+ var cpuUtil float64
+ if vm.Runtime.MaxCpuUsage != 0 {
+ cpuUtil = 100 * float64(cpuUsage) / float64(vm.Runtime.MaxCpuUsage)
+ } else {
+ cpuUtil = 100 * float64(cpuUsage) / float64(ncpu*hs.Summary.Hardware.CpuMhz)
+ }
+
+ v.mb.RecordVcenterVMCPUUsageDataPoint(now, int64(cpuUsage))
+ v.mb.RecordVcenterVMCPUUtilizationDataPoint(now, cpuUtil)
+
 }
 
 func (v *vcenterMetricScraper) recordDatastoreProperties(