[receiver/vcenterreceiver] Optimize vCenter VM network calls (open-te…

…lemetry#32201)

**Description:** <Describe what has changed.>
<!--Ex. Fixing a bug - Describe the bug and how this fixes the issue.
Ex. Adding a feature - Explain what this achieves.-->
Changes the method for collecting VMs used by the `vccenterreceiver` to
the more time-efficient `CreateContainerView` method. This is the first
step to addressing the issue linked below.

**Link to tracking Issue:** open-telemetry#31837

**Testing:** These changes were tested on an environment with 200+
virtual machines. The original collection time was ~80 seconds.
Collection times with these changes are ~40 seconds.

**Documentation:** N/A

---------

Co-authored-by: Stefan Kurek <[email protected]>

.chloggen/feat_faster-vcenter-vm-network-calls.yaml

@@ -0,0 +1,27 @@
+# Use this changelog template to create an entry for release notes.
+
+# 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: "Changes process for collecting VMs & VM perf metrics used by the `vccenterreceiver` to be more efficient (one call now for all VMs)"
+
+# Mandatory: One or more tracking issues related to the change. You can use the PR number here if no issue exists.
+issues: [31837]
+
+# (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:
+
+# If your change doesn't affect end users or the exported elements of any package,
+# you should instead start your pull request title with [chore] or use the "Skip Changelog" label.
+# Optional: The change log or logs in which this entry should be included.
+# e.g. '[user]' or '[user, api]'
+# Include 'user' if the change is relevant to end users.
+# Include 'api' if there is a change to a library API.
+# Default: '[user]'
+change_logs: [user]

receiver/vcenterreceiver/client.go

@@ -13,7 +13,9 @@ import (
 	"github.com/vmware/govmomi/object"
 	"github.com/vmware/govmomi/performance"
 	"github.com/vmware/govmomi/property"
+	"github.com/vmware/govmomi/view"
 	"github.com/vmware/govmomi/vim25"
+	"github.com/vmware/govmomi/vim25/mo"
 	vt "github.com/vmware/govmomi/vim25/types"
 )
 
@@ -24,6 +26,7 @@ type vcenterClient struct {
 	finder    *find.Finder
 	pc        *property.Collector
 	pm        *performance.Manager
+	vm        *view.Manager
 	cfg       *Config
 }
 
@@ -69,6 +72,7 @@ func (vc *vcenterClient) EnsureConnection(ctx context.Context) error {
 	vc.pc = property.DefaultCollector(vc.vimDriver)
 	vc.finder = find.NewFinder(vc.vimDriver)
 	vc.pm = performance.NewManager(vc.vimDriver)
+	vc.vm = view.NewManager(vc.vimDriver)
 	return nil
 }
 
@@ -89,6 +93,7 @@ func (vc *vcenterClient) Datacenters(ctx context.Context) ([]*object.Datacenter,
 	return datacenters, nil
 }
 
+// Computes returns the ComputeResources (and ClusterComputeResources) of the vSphere SDK for a given datacenter
 func (vc *vcenterClient) Computes(ctx context.Context, datacenter *object.Datacenter) ([]*object.ComputeResource, error) {
 	vc.finder = vc.finder.SetDatacenter(datacenter)
 	computes, err := vc.finder.ComputeResourceList(ctx, "*")
@@ -98,7 +103,7 @@ func (vc *vcenterClient) Computes(ctx context.Context, datacenter *object.Datace
 	return computes, nil
 }
 
-// ResourcePools returns the resourcePools in the vSphere SDK
+// ResourcePools returns the ResourcePools in the vSphere SDK
 func (vc *vcenterClient) ResourcePools(ctx context.Context) ([]*object.ResourcePool, error) {
 	rps, err := vc.finder.ResourcePoolList(ctx, "*")
 	if err != nil {
@@ -107,19 +112,47 @@ func (vc *vcenterClient) ResourcePools(ctx context.Context) ([]*object.ResourceP
 	return rps, err
 }
 
-func (vc *vcenterClient) VMs(ctx context.Context) ([]*object.VirtualMachine, error) {
-	vms, err := vc.finder.VirtualMachineList(ctx, "*")
+func (vc *vcenterClient) VMs(ctx context.Context) ([]mo.VirtualMachine, error) {
+	v, err := vc.vm.CreateContainerView(ctx, vc.vimDriver.ServiceContent.RootFolder, []string{"VirtualMachine"}, true)
 	if err != nil {
-		return nil, fmt.Errorf("unable to retrieve vms: %w", err)
+		return nil, fmt.Errorf("unable to retrieve VMs: %w", err)
 	}
-	return vms, err
+
+	var vms []mo.VirtualMachine
+	err = v.Retrieve(ctx, []string{"VirtualMachine"}, []string{
+		"config.hardware.numCPU",
+		"config.instanceUuid",
+		"runtime.powerState",
+		"runtime.maxCpuUsage",
+		"summary.quickStats.guestMemoryUsage",
+		"summary.quickStats.balloonedMemory",
+		"summary.quickStats.swappedMemory",
+		"summary.quickStats.ssdSwappedMemory",
+		"summary.quickStats.overallCpuUsage",
+		"summary.config.memorySizeMB",
+		"summary.config.name",
+		"summary.storage.committed",
+		"summary.storage.uncommitted",
+		"summary.runtime.host",
+		"resourcePool",
+	}, &vms)
+	if err != nil {
+		return nil, fmt.Errorf("unable to retrieve VMs: %w", err)
+	}
+
+	return vms, nil
 }
 
 type perfSampleResult struct {
 	counters map[string]*vt.PerfCounterInfo
 	results  []performance.EntityMetric
 }
 
+type perfMetricsQueryResult struct {
+	counters       map[string]*vt.PerfCounterInfo
+	resultsByMoRef map[string]*performance.EntityMetric
+}
+
 func (vc *vcenterClient) performanceQuery(
 	ctx context.Context,
 	spec vt.PerfQuerySpec,
@@ -147,3 +180,36 @@ func (vc *vcenterClient) performanceQuery(
 		results:  result,
 	}, nil
 }
+
+func (vc *vcenterClient) perfMetricsQuery(
+	ctx context.Context,
+	spec vt.PerfQuerySpec,
+	names []string,
+	objs []vt.ManagedObjectReference,
+) (*perfMetricsQueryResult, error) {
+	if vc.pm == nil {
+		return &perfMetricsQueryResult{}, nil
+	}
+	vc.pm.Sort = true
+	sample, err := vc.pm.SampleByName(ctx, spec, names, objs)
+	if err != nil {
+		return nil, err
+	}
+	result, err := vc.pm.ToMetricSeries(ctx, sample)
+	if err != nil {
+		return nil, err
+	}
+	counterInfoByName, err := vc.pm.CounterInfoByName(ctx)
+	if err != nil {
+		return nil, err
+	}
+
+	resultsByMoRef := map[string]*performance.EntityMetric{}
+	for i := range result {
+		resultsByMoRef[result[i].Entity.Value] = &result[i]
+	}
+	return &perfMetricsQueryResult{
+		counters:       counterInfoByName,
+		resultsByMoRef: resultsByMoRef,
+	}, nil
+}

receiver/vcenterreceiver/client_test.go

@@ -13,6 +13,7 @@ import (
 	"github.com/vmware/govmomi/find"
 	"github.com/vmware/govmomi/session"
 	"github.com/vmware/govmomi/simulator"
+	"github.com/vmware/govmomi/view"
 	"github.com/vmware/govmomi/vim25"
 	"go.opentelemetry.io/collector/config/configopaque"
 	"go.opentelemetry.io/collector/config/configtls"
@@ -48,10 +49,12 @@ func TestGetResourcePools(t *testing.T) {
 
 func TestGetVMs(t *testing.T) {
 	simulator.Test(func(ctx context.Context, c *vim25.Client) {
+		viewManager := view.NewManager(c)
 		finder := find.NewFinder(c)
 		client := vcenterClient{
 			vimDriver: c,
 			finder:    finder,
+			vm:        viewManager,
 		}
 		vms, err := client.VMs(ctx)
 		require.NoError(t, err)

receiver/vcenterreceiver/internal/mockserver/client_mock.go

@@ -79,6 +79,8 @@ func routeBody(t *testing.T, requestType string, body map[string]any) ([]byte, e
 		return routeRetreiveProperties(t, body)
 	case "QueryPerf":
 		return routePerformanceQuery(t, body)
+	case "CreateContainerView":
+		return loadResponse("create-container-view.xml")
 	}
 
 	return []byte{}, errNotFound
@@ -163,7 +165,7 @@ func routeRetreiveProperties(t *testing.T, body map[string]any) ([]byte, error)
 	case contentType == "HostSystem":
 		if ps, ok := propSet["pathSet"].([]any); ok {
 			for _, v := range ps {
-				if v == "summary.hardware" {
+				if v == "summary.hardware" || v == "summary.hardware.cpuMhz" {
 					return loadResponse("host-properties.xml")
 				}
 			}
@@ -179,26 +181,8 @@ func routeRetreiveProperties(t *testing.T, body map[string]any) ([]byte, error)
 
 		}
 
-	case content == "group-v4" && contentType == "Folder":
-		if propSetArray {
-			return loadResponse("vm-group.xml")
-		}
-		if propSet == nil {
-			return loadResponse("vm-folder.xml")
-		}
-		return loadResponse("vm-folder-parent.xml")
-
-	case content == "vm-1040" && contentType == "VirtualMachine":
-		if propSet["pathSet"] == "summary.runtime.host" {
-			return loadResponse("vm-host.xml")
-		}
-		if propSet["pathSet"] == "resourcePool" {
-			return loadResponse("vm-resource-pool.xml")
-		}
-		return loadResponse("vm-properties.xml")
-
-	case (content == "group-v1034" || content == "group-v1001") && contentType == "Folder":
-		return loadResponse("vm-empty-folder.xml")
+	case contentType == "ContainerView" && propSet["type"] == "VirtualMachine":
+		return loadResponse("vm-default-properties.xml")
 
 	case contentType == "ResourcePool":
 		if ps, ok := propSet["pathSet"].([]any); ok {
@@ -238,7 +222,11 @@ func routeRetreiveProperties(t *testing.T, body map[string]any) ([]byte, error)
 func routePerformanceQuery(t *testing.T, body map[string]any) ([]byte, error) {
 	queryPerf := body["QueryPerf"].(map[string]any)
 	require.NotNil(t, queryPerf)
-	querySpec := queryPerf["querySpec"].(map[string]any)
+	querySpec, ok := queryPerf["querySpec"].(map[string]any)
+	if !ok {
+		querySpecs := queryPerf["querySpec"].([]any)
+		querySpec = querySpecs[0].(map[string]any)
+	}
 	entity := querySpec["entity"].(map[string]any)
 	switch entity["-type"] {
 	case "HostSystem":

receiver/vcenterreceiver/internal/mockserver/responses/vm-empty-folder.xml → receiver/vcenterreceiver/internal/mockserver/responses/create-container-view.xml

@@ -1,6 +1,8 @@
 <?xml version="1.0" encoding="UTF-8"?>
 <soapenv:Envelope xmlns:soapenc="http://schemas.xmlsoap.org/soap/encoding/" xmlns:soapenv="http://schemas.xmlsoap.org/soap/envelope/" xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
     <soapenv:Body>
-        <RetrievePropertiesResponse xmlns="urn:vim25"></RetrievePropertiesResponse>
+        <CreateContainerViewResponse xmlns="urn:vim25">
+            <returnval type="ContainerView">session[5279c748-1a9c-e5f0-cc05-71b15da3a8a6]529d3b1c-0a03-30d0-f0c8-8cae2cc07d25</returnval>
+        </CreateContainerViewResponse>
     </soapenv:Body>
-</soapenv:Envelope>
+</soapenv:Envelope>

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

@@ -4,6 +4,14 @@
         <RetrievePropertiesResponse xmlns="urn:vim25">
             <returnval>
                 <obj type="HostSystem">host-1002</obj>
+                <propSet>
+                    <name>name</name>
+                    <val xsi:type="xsd:string">esxi-27971.cf5e88ac.australia-southeast1.gve.goog</val>
+                </propSet>
+                <propSet>
+                    <name>summary.hardware.cpuMhz</name>
+                    <val xsi:type="xsd:int">2593</val>
+                </propSet>
                 <propSet>
                     <name>config</name>
                     <val xsi:type="HostConfigInfo">

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

@@ -0,0 +1,70 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<soapenv:Envelope xmlns:soapenc="http://schemas.xmlsoap.org/soap/encoding/" xmlns:soapenv="http://schemas.xmlsoap.org/soap/envelope/" xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
+    <soapenv:Body>
+        <RetrievePropertiesResponse xmlns="urn:vim25">
+            <returnval>
+                <obj type="VirtualMachine">vm-1040</obj>
+                <propSet>
+                    <name>config.hardware.numCPU</name>
+                    <val xsi:type="xsd:int">4</val>
+                </propSet>
+                <propSet>
+                    <name>config.instanceUuid</name>
+                    <val xsi:type="xsd:string">5000bbe0-993e-5813-c56a-198eaa62fb61</val>
+                </propSet>
+                <propSet>
+                    <name>resourcePool</name>
+                    <val type="ResourcePool" xsi:type="ManagedObjectReference">resgroup-9</val>
+                </propSet>
+                <propSet>
+                    <name>runtime.maxCpuUsage</name>
+                    <val xsi:type="xsd:int">10372</val>
+                </propSet>
+                <propSet>
+                    <name>runtime.powerState</name>
+                    <val xsi:type="VirtualMachinePowerState">poweredOn</val>
+                </propSet>
+                <propSet>
+                    <name>summary.config.memorySizeMB</name>
+                    <val xsi:type="xsd:int">16384</val>
+                </propSet>
+                <propSet>
+                    <name>summary.config.name</name>
+                    <val xsi:type="xsd:string">CentOS 7</val>
+                </propSet>
+                <propSet>
+                    <name>summary.quickStats.balloonedMemory</name>
+                    <val xsi:type="xsd:int">0</val>
+                </propSet>
+                <propSet>
+                    <name>summary.quickStats.guestMemoryUsage</name>
+                    <val xsi:type="xsd:int">163</val>
+                </propSet>
+                <propSet>
+                    <name>summary.quickStats.overallCpuUsage</name>
+                    <val xsi:type="xsd:int">12</val>
+                </propSet>
+                <propSet>
+                    <name>summary.quickStats.ssdSwappedMemory</name>
+                    <val xsi:type="xsd:long">0</val>
+                </propSet>
+                <propSet>
+                    <name>summary.quickStats.swappedMemory</name>
+                    <val xsi:type="xsd:int">0</val>
+                </propSet>
+                <propSet>
+                    <name>summary.runtime.host</name>
+                    <val type="HostSystem" xsi:type="ManagedObjectReference">host-1002</val>
+                </propSet>
+                <propSet>
+                    <name>summary.storage.committed</name>
+                    <val xsi:type="xsd:long">16311648256</val>
+                </propSet>
+                <propSet>
+                    <name>summary.storage.uncommitted</name>
+                    <val xsi:type="xsd:long">258847277056</val>
+                </propSet>
+            </returnval>
+        </RetrievePropertiesResponse>
+    </soapenv:Body>
+</soapenv:Envelope>