Skip to content

Commit

Permalink
feat(inputs.vsphere): Collect resource pools metrics and add resource…
Browse files Browse the repository at this point in the history 
… pool tag in VM metrics (#10574)

* Collect Resource Pool metrics; Add rpname tag on VM metrics

* Update vSphere readme file

* Update vSphere readme file

* Correct typo in vSphere Readme

* Correct Markdown of metrics.md

* Fix metrics file

* Fix code in endpoint (filter); add some tests

* Update plugins/inputs/vsphere/endpoint.go

That's true I commit this suggestion

Co-authored-by: Sebastian Spaink <[email protected]>

* Removed Context and Endpoint from getResourcePoolName func

Co-authored-by: Simon LAMBERT <[email protected]>
Co-authored-by: Sebastian Spaink <[email protected]>
  • Loading branch information
@6monlambert @sspaink @MyaLongmire
3 people authored and MyaLongmire committed Jul 6, 2022
1 parent 0e9b2ee commit 9c840ed
Show file tree
Hide file tree
Showing 6 changed files with 233 additions and 71 deletions.
53 changes: 53 additions & 0 deletions plugins/inputs/vsphere/METRICS.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -193,6 +193,59 @@ vmop.numSVMotion.latest
vmop.numXVMotion.latest
```

## Resource Pool Metrics

```metrics
cpu.usagemhz.average
cpu.cpuentitlement.latest
cpu.usagemhz.minimum
cpu.usagemhz.maximum
cpu.capacity.entitlement.average
cpu.capacity.usage.average
cpu.capacity.demand.average
cpu.capacity.contention.average
cpu.corecount.provisioned.average
cpu.corecount.contention.average
disk.throughput.usage.average
disk.throughput.contention.average
mem.capacity.contention.average
mem.overhead.average
mem.consumed.average
mem.granted.average
mem.active.average
mem.shared.average
mem.zero.average
mem.swapped.average
mem.vmmemctl.average
mem.capacity.provisioned.average
mem.capacity.entitlement.average
mem.capacity.usage.average
mem.mementitlement.latest
mem.compressed.average
mem.compressionRate.average
mem.decompressionRate.average
mem.overhead.minimum
mem.consumed.minimum
mem.granted.minimum
mem.active.minimum
mem.shared.minimum
mem.zero.minimum
mem.swapped.minimum
mem.vmmemctl.maximum
mem.overhead.maximum
mem.consumed.maximum
mem.granted.maximum
mem.active.maximum
mem.shared.maximum
mem.zero.maximum
mem.swapped.maximum
mem.vmmemctl.minimum
net.throughput.usage.average
net.throughput.contention.summation
power.power.average
power.energy.summation
```

## Cluster Metrics

```metrics
Expand Down
27 changes: 23 additions & 4 deletions plugins/inputs/vsphere/README.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -4,6 +4,7 @@ The VMware vSphere plugin uses the vSphere API to gather metrics from multiple v

* Clusters
* Hosts
* Resource Pools
* VMs
* Datastores

Expand Down Expand Up @@ -140,7 +141,14 @@ vm_metric_exclude = [ "*" ]
# cluster_metric_exclude = [] ## Nothing excluded by default
# cluster_instances = false ## false by default

## Datastores
## Resource Pools
# datastore_include = [ "/*/host/**"] # Inventory path to datastores to collect (by default all are collected)
# datastore_exclude = [] # Inventory paths to exclude
# datastore_metric_include = [] ## if omitted or empty, all metrics are collected
# datastore_metric_exclude = [] ## Nothing excluded by default
# datastore_instances = false ## false by default

## Datastores
# datastore_include = [ "/*/datastore/**"] # Inventory path to datastores to collect (by default all are collected)
# datastore_exclude = [] # Inventory paths to exclude
# datastore_metric_include = [] ## if omitted or empty, all metrics are collected
Expand Down Expand Up @@ -252,10 +260,13 @@ to a file system. A vSphere inventory has a structure similar to this:
| | | +-VM1
| | | +-VM2
| | | +-hadoop1
| +-Host2 # Dummy cluster created for non-clustered host
| | +-Host2
| | +-ResourcePool1
| | | +-VM3
| | | +-VM4
| +-Host2 # Dummy cluster created for non-clustered host
| | +-Host2
| | | +-VM5
| | | +-VM6
+-vm # VM folder (created by system)
| +-VM1
| +-VM2
Expand Down Expand Up @@ -289,7 +300,7 @@ We can extend this to looking at a cluster level: ```/DC0/host/Cluster1/*/hadoop
vCenter keeps two different kinds of metrics, known as realtime and historical metrics.

* Realtime metrics: Available at a 20 second granularity. These metrics are stored in memory and are very fast and cheap to query. Our tests have shown that a complete set of realtime metrics for 7000 virtual machines can be obtained in less than 20 seconds. Realtime metrics are only available on **ESXi hosts** and **virtual machine** resources. Realtime metrics are only stored for 1 hour in vCenter.
* Historical metrics: Available at a (default) 5 minute, 30 minutes, 2 hours and 24 hours rollup levels. The vSphere Telegraf plugin only uses the most granular rollup which defaults to 5 minutes but can be changed in vCenter to other interval durations. These metrics are stored in the vCenter database and can be expensive and slow to query. Historical metrics are the only type of metrics available for **clusters**, **datastores** and **datacenters**.
* Historical metrics: Available at a (default) 5 minute, 30 minutes, 2 hours and 24 hours rollup levels. The vSphere Telegraf plugin only uses the most granular rollup which defaults to 5 minutes but can be changed in vCenter to other interval durations. These metrics are stored in the vCenter database and can be expensive and slow to query. Historical metrics are the only type of metrics available for **clusters**, **datastores**, **resource pools** and **datacenters**.

For more information, refer to the vSphere documentation here: <https://pubs.vmware.com/vsphere-50/index.jsp?topic=%2Fcom.vmware.wssdk.pg.doc_50%2FPG_Ch16_Performance.18.2.html>

Expand All @@ -314,6 +325,7 @@ This will disrupt the metric collection and can result in missed samples. The be
datastore_metric_exclude = ["*"]
cluster_metric_exclude = ["*"]
datacenter_metric_exclude = ["*"]
resourcepool_metric_exclude = ["*"]

collect_concurrency = 5
discover_concurrency = 5
Expand Down Expand Up @@ -400,6 +412,12 @@ When the vSphere plugin queries vCenter for historical statistics it queries for
* Res CPU: active, max, running
* System: operating system uptime, uptime
* Virtual Disk: seeks, # reads/writes, latency, load
* Resource Pools stats:
* Memory: total, usage, active, latency, swap, shared, vmmemctl
* CPU: capacity, usage, corecount
* Disk: throughput
* Network: throughput
* Power: energy, usage
* Datastore stats:
* Disk: Capacity, provisioned, used

Expand All @@ -415,6 +433,7 @@ For a detailed list of commonly available metrics, please refer to [METRICS.md](
* cluster (vcenter cluster)
* esxhost (name of ESXi host)
* guest (guest operating system id)
* resource pool (name of resource pool)
* cpu stats for Host and VM
* cpu (cpu core - not all CPU fields will have this tag)
* datastore stats for Host and VM
Expand Down
68 changes: 68 additions & 0 deletions plugins/inputs/vsphere/endpoint.go
View file
Original file line number Diff line number Diff line change
Expand Up @@ -100,6 +100,7 @@ type objectRef struct {
parentRef *types.ManagedObjectReference //Pointer because it must be nillable
guest string
dcname string
rpname string
customValues map[string]string
lookup map[string]string
}
Expand Down Expand Up @@ -165,6 +166,24 @@ func NewEndpoint(ctx context.Context, parent *VSphere, address *url.URL, log tel
getObjects: getClusters,
parent: "datacenter",
},
"resourcepool": {
name: "resourcepool",
vcName: "ResourcePool",
pKey: "rpname",
parentTag: "clustername",
enabled: anythingEnabled(parent.ResourcePoolMetricExclude),
realTime: false,
sampling: int32(time.Duration(parent.HistoricalInterval).Seconds()),
objects: make(objectMap),
filters: newFilterOrPanic(parent.ResourcePoolMetricInclude, parent.ResourcePoolMetricExclude),
paths: parent.ResourcePoolInclude,
excludePaths: parent.ResourcePoolExclude,
simple: isSimple(parent.ResourcePoolMetricInclude, parent.ResourcePoolMetricExclude),
include: parent.ResourcePoolMetricInclude,
collectInstances: parent.ResourcePoolInstances,
getObjects: getResourcePools,
parent: "cluster",
},
"host": {
name: "host",
vcName: "HostSystem",
Expand Down Expand Up @@ -653,6 +672,35 @@ func getClusters(ctx context.Context, e *Endpoint, resourceFilter *ResourceFilte
return m, nil
}

//noinspection GoUnusedParameter
func getResourcePools(ctx context.Context, e *Endpoint, resourceFilter *ResourceFilter) (objectMap, error) {
var resources []mo.ResourcePool
err := resourceFilter.FindAll(ctx, &resources)
if err != nil {
return nil, err
}
m := make(objectMap)
for _, r := range resources {
m[r.ExtensibleManagedObject.Reference().Value] = &objectRef{
name: r.Name,
ref: r.ExtensibleManagedObject.Reference(),
parentRef: r.Parent,
customValues: e.loadCustomAttributes(&r.ManagedEntity),
}
}
return m, nil
}

func getResourcePoolName(rp types.ManagedObjectReference, rps objectMap) string {
//Loop through the Resource Pools objectmap to find the corresponding one
for _, r := range rps {
if r.ref == rp {
return r.name
}
}
return "Resources" //Default value
}

//noinspection GoUnusedParameter
func getHosts(ctx context.Context, e *Endpoint, resourceFilter *ResourceFilter) (objectMap, error) {
var resources []mo.HostSystem
Expand Down Expand Up @@ -681,13 +729,29 @@ func getVMs(ctx context.Context, e *Endpoint, resourceFilter *ResourceFilter) (o
return nil, err
}
m := make(objectMap)
client, err := e.clientFactory.GetClient(ctx)
if err != nil {
return nil, err
}
//Create a ResourcePool Filter and get the list of Resource Pools
rprf := ResourceFilter{
finder: &Finder{client},
resType: "ResourcePool",
paths: []string{"/*/host/**"},
excludePaths: nil}
resourcePools, err := getResourcePools(ctx, e, &rprf)
if err != nil {
return nil, err
}
for _, r := range resources {
if r.Runtime.PowerState != "poweredOn" {
continue
}
guest := "unknown"
uuid := ""
lookup := make(map[string]string)
// Get the name of the VM resource pool
rpname := getResourcePoolName(*r.ResourcePool, resourcePools)

// Extract host name
if r.Guest != nil && r.Guest.HostName != "" {
Expand Down Expand Up @@ -755,6 +819,7 @@ func getVMs(ctx context.Context, e *Endpoint, resourceFilter *ResourceFilter) (o
parentRef: r.Runtime.Host,
guest: guest,
altID: uuid,
rpname: rpname,
customValues: e.loadCustomAttributes(&r.ManagedEntity),
lookup: lookup,
}
Expand Down Expand Up @@ -1191,6 +1256,9 @@ func (e *Endpoint) populateTags(objectRef *objectRef, resourceType string, resou
if resourceType == "vm" && objectRef.altID != "" {
t["uuid"] = objectRef.altID
}
if resourceType == "vm" && objectRef.rpname != "" {
t["rpname"] = objectRef.rpname
}

// Map parent reference
parent, found := e.getParent(objectRef, resource)
Expand Down
6 changes: 4 additions & 2 deletions plugins/inputs/vsphere/finder.go
View file
Original file line number Diff line number Diff line change
Expand Up @@ -246,6 +246,7 @@ func matchName(f property.Filter, props []types.DynamicProperty) bool {
func init() {
childTypes = map[string][]string{
"HostSystem": {"VirtualMachine"},
"ResourcePool": {"VirtualMachine"},
"ComputeResource": {"HostSystem", "ResourcePool", "VirtualApp"},
"ClusterComputeResource": {"HostSystem", "ResourcePool", "VirtualApp"},
"Datacenter": {"Folder"},
Expand All @@ -260,9 +261,10 @@ func init() {
}

addFields = map[string][]string{
"HostSystem": {"parent", "summary.customValue", "customValue"},
"HostSystem": {"parent", "summary.customValue", "customValue"},
"ResourcePool": {"parent", "customValue"},
"VirtualMachine": {"runtime.host", "config.guestId", "config.uuid", "runtime.powerState",
"summary.customValue", "guest.net", "guest.hostName", "customValue"},
"summary.customValue", "guest.net", "guest.hostName", "resourcePool", "customValue"},
"Datastore": {"parent", "info", "customValue"},
"ClusterComputeResource": {"parent", "customValue"},
"Datacenter": {"parent", "customValue"},
Expand Down
Loading

0 comments on commit 9c840ed

Please sign in to comment.