VMware vSphere Monitoring Overview

Starter

Professional

Prerequisites

vSphere monitoring architecture: Glouton agent on a host machine connecting to vCenter API to monitor ESXi hosts and VMs, sending metrics to Bleemeo Cloud

VMware vSphere monitoring requires a Bleemeo Glouton agent installed and connected to your Bleemeo account on a Linux or Windows machine. The agent communicates with the vSphere API to retrieve facts and metrics from clusters, hosts, and VMs.

Glouton supports both vCenter and standalone ESXi hosts. When connecting to a vCenter, all managed hosts, VMs, and datastores are automatically discovered.

The vSphere user account needs read-only permissions on the objects you want to monitor (datacenter, cluster, hosts, VMs).

See the installation guide if you don’t have Glouton installed yet.

Configuration

To enable VMware vSphere monitoring, add the vCenter or ESXi connection details to your Glouton configuration. You can monitor multiple vCenter appliances or standalone ESXi hosts.

Linux
Windows

sudo tee /etc/glouton/conf.d/99-vsphere.conf > /dev/null << 'EOF'
vsphere:
  - url: "https://<vcenter-url>/sdk"
    username: "<username>"
    password: "<password>"
    insecure_skip_verify: false
    skip_monitor_vms: false
EOF

sudo systemctl restart glouton

Create or edit the file C:\ProgramData\glouton\conf.d\99-vsphere.conf:

vsphere:
  - url: "https://<vcenter-url>/sdk"
    username: "<username>"
    password: "<password>"
    insecure_skip_verify: false
    skip_monitor_vms: false

Then restart the Glouton service:

Restart-Service glouton

Configuration Options

Option	Description
`url`	URL of the vCenter or ESXi SDK endpoint (e.g., `https://vcenter.example.com/sdk`)
`username`	Username for vCenter authentication
`password`	Password for vCenter authentication
`insecure_skip_verify`	Set to `true` to skip SSL certificate verification (use only for self-signed certificates)
`skip_monitor_vms`	Set to `true` to only monitor hosts and skip VM metrics

Metrics

Host

Metric	Description
`cpu_used`	Percentage of CPU capacity being used by the host
`mem_total`	Total reservation, available and consumed, for powered-on VMs on the host
`mem_used_perc`	Percentage of memory being used by the host
`swap_in`	Amount of memory swapped in by the host
`swap_out`	Amount of memory swapped out by the host
`io_read_bytes`	Amount of data read from the disk by the host
`io_write_bytes`	Amount of data written to the disk by the host
`net_bits_sent`	Average rate at which data was transmitted by the host
`net_bits_recv`	Average rate at which data was received by the host
`vms_running_count`	Count of VM(s) currently running on the host
`vms_stopped_count`	Count of VM(s) currently stopped on the host

VM

Metric	Description
`cpu_used`	CPU utilization by the VM
`vsphere_vm_cpu_latency_perc`	Percent of the time the VM is waiting for CPU to be available
`mem_used_perc`	Percent of the memory used by the VM
`swap_used`	Amount of swapped memory of the VM
`io_read_bytes`	Amount read from the disk by the VM
`io_write_bytes`	Amount written to the disk by the VM
`net_bits_sent`	Data transmission rate of the VM
`net_bits_recv`	Data reception rate of the VM
`disk_used_perc`	Filesystem space used in percent

Labels

vSphere metrics have the following labels:

dcname
clustername (not for standalone ESXis)
esxhostname (only for hosts and VMs)
vmname (only for VMs)

Additionally, metrics related to datastore, network or disk have an item label, which represents:

a specific datastore
a network interface
a disk
a partition