We are experiencing inconsistent memory usage values between VMware-based monitoring and OS-level monitoring in Zabbix.

Issue Details:

• VMs discovered via the VMware hypervisor show a memory usage value that does not match the real usage inside the guest OS.

• Zabbix Agent/Agent2 running inside the VM reports significantly different memory statistics.

• “Active Memory” metrics (e.g., from VMware performance counters) cannot be collected — Zabbix either returns unsupported items or does not expose the metric at all.

• The mismatch creates confusion in dashboards and alerting because VMware-based and OS-based metrics contradict each other.

Environment:

• Zabbix Server: 7.0.19

• VMware vCenter version: (fill this)

• Zabbix VMware Discovery using template: Template VM VMware Hypervisor

• OS Agents: Zabbix Agent/Agent2 on Windows/Linux (fill)

Steps to Reproduce:

1. Discover VMs through VMware in Zabbix.

1. Compare:

• • vmware.vm.memory.usage

• • OS-level memory usage from agent (e.g., vm.memory.size[available], vm.memory.size[total], etc.)

1. Attempt to collect active memory metrics (e.g., mem.active.average)

1. Observe the mismatch and missing values.

Expected Result:

• Zabbix should report memory usage consistently between VMware and OS-level metrics.

• VMware Active Memory should be available as an item or have a clear documented limitation.

Actual Result:

• Memory utilization differs significantly between VMware and OS-level values.

• Active Memory values are not available / cannot be collected.