I manage a couple of enterprise cloud enviroments and I recently came across this error on one of the VMware hosts, commonly referred to as the VMware purple screen of death (PSOD)!
After doing some digging all of the templates for the VMs use the E1000 NIC. It maybe worth checking your NICs to see whether you are also using E1000. Also if the operating system is using RSS (Receive Side Scaling which is being used by pretty much all operating systems since Windows Vista) it is possible that your ESXi 5.0/5.1/5.5 host will shell out with a PSOD (Purple Screen of Death) with errors containing E1000PollRxRing.
The immediate resolution is to patch your hosts, links to VMware KBs below.
1. See KB2062300 for more information / Patch from KB 2062300 is included in ESXi 5.1 Update 2
2. The E1000 issue is fixed in VMware ESXi 5.5, Patch Release ESXi550-201410001 (2087358) VMware ESXi 5.5, Patch Release ESXi550-201410001.
Now like all enviroments arranging downtime to patch hosts can be tricky, but there are a couple of workarounds.
1. Stop using the E1000 NICs and use the VMXNET3 NIC instead
2. If you really want to you can diable RSS (as mentioned above) using the following command. Note this is only for servers Win 2008 +.
netsh int tcp set global rss=disabled