Applications Systems

VMware ESXi Host NICs Show Impossibly High Network Utilization in SolarWinds (Dell PowerEdge, ESXi 8.0.3)

ESXi hosts on Dell PowerEdge hardware running ESXi 8.0.3 Build 25205845 report impossibly high vmnic NetworkUtilization values in SolarWinds (e.g., 1,000%–14,000%) that do not match actual physical NIC traffic or vCenter performance charts. This is caused by incorrect network performance counter values returned by the vSphere API for these specific hosts. SolarWinds has no validation layer for VMware SDK metrics and stores and displays whatever the API returns. The issue must be resolved at the ESXi/vSphere level with Broadcom support.

First published date

7/22/2026 10:36 PM

Last published date

7/22/2026 10:36 PM

Overview

Issue

SolarWinds reports impossibly high network utilization values on vmnic interfaces for ESXi hosts — for example, Receive values of 1+ Tbps on a 25 GbE NIC, with Percent Utilization showing 1,000% to 14,000% or higher. The values displayed in SolarWinds do not match the physical NIC traffic, switch port statistics, or performance counters visible in vCenter for the same hosts at the same timestamps.

The following characteristics are common across reported cases:

  • Affected hosts are Dell Inc. hardware, specifically PowerEdge R650, PowerEdge R660, or R660 vSAN Ready Node models

  • ESXi Version: 8.0.3, Build 25205845

  • The issue affects multiple hosts across multiple clusters simultaneously — not isolated to a single host

  • The issue is present on both managed SolarWinds Platform nodes and VMware-discovered-only hosts (NodeID = NULL)

  • SolarWinds VMware polling completes successfully with no errors in VIM.Pollers.Jobs logs — the inflated values are what the VMware API is returning

Product section

Virtualization Manager

Cause

SolarWinds monitors ESXi host network utilization via the VMware SDK using the vSphere API — not via SNMP. For affected Dell PowerEdge hosts running ESXi 8.0.3 Build 25205845, the network performance counters returned by the vSphere API contain values that are significantly above the actual NIC rate.

SolarWinds does not apply counter validation boundaries to VMware SDK metrics. Values returned from the API are stored and displayed as-is. The inflated values originate at the VMware API layer — there is no correction that can be applied on the SolarWinds side without the underlying counter being fixed at the ESXi level.

Resolution

Resolution

  • Confirm the issue in vSphere — check the network receive counter for affected vmnic interfaces directly in vCenter Performance charts at the same timestamps reported in SolarWinds. If vCenter also shows inflated or anomalous values, the counter issue is confirmed at the ESXi layer.

  • Open a support case with Broadcom (VMware support) referencing the affected ESXi build (8.0.3, Build 25205845) and the inflated network counter behavior on Dell PowerEdge R650/R660 hardware.

  • There is no SolarWinds-side fix or workaround for this issue. The problem will persist until the network performance counter is corrected at the ESXi/vSphere level.

Additional Notes

  • This issue has been observed across multiple environments on Dell PowerEdge hardware running ESXi 8.0.3.

  • VIM.Pollers.Jobs logs will show no WARN or ERROR entries during the affected period — polling is completing successfully, and the inflated values are present in the API response itself.

  • The issue affects the NetworkUtilization and NetworkReceiveRate columns in VIM_Hosts.